Recursive Genome Function of theCerebellum: Geometric Unification ofNeuroscience and Genomics
61
Andras J. Pellionisz, Roy Graham, Peter A. Pellionisz, andJean-Claude Perez
Abstract
Recursive Fractal Genome Function in the geometric mind frame of Tensor
Network Theory (TNT) leads through FractoGene to a mathematical unification
of physiological and pathological development of neural structure and function
as governed by the genome. The cerebellum serves as the best platform for
unification of neuroscience and genomics. The matrix of massively parallel
neural nets of fractal Purkinje brain cells explains the sensorimotor,
multidimensional non-Euclidean coordination by the cerebellum acting as
a space-time metric tensor. In TNT, the recursion of covariant sensory vectors
into contravariant motor executions converges into Eigenstates composing the
cerebellar metric as a Moore-Penrose Pseudo-Inverse.
The Principle of Recursion is generalized to genomic systems with the
realization that the assembly of proteins from nucleic acids as governed by
regulation of coding RNA (cRNA) is a contravariant multicomponent functor,
where in turn the quantum states of resulting protein structures both in intergenic
and intronic sequences are measured in a covariant manner by noncoding RNA
(ncRNA) arising as a result of proteins binding with ncDNA modulated by
transcription factors. Thus, cRNA and ncRNA vectors by their interference
constitute a genomic metric, the RNA system serving as a Genomic
A.J. Pellionisz (*)
HolGenTech, Sunnyvale, CA, 94086, USA
e-mail: [email protected], [email protected]
R. Graham
DRC Computer, Sunnyvale, CA, 94089, USA
P.A. Pellionisz
UCLA, Westwood, CA, 90024, USA
J.-C. Perez
IBM Emeritus, Martignas, 33127, France
e-mail: [email protected]
M. Manto, D.L. Gruol, J.D. Schmahmann, N. Koibuchi, F. Rossi (eds.),
Handbook of the Cerebellum and Cerebellar Disorders,DOI 10.1007/978-94-007-1333-8_61, # Springer Science+Business Media Dordrecht 2013
1381
Cerebellum. Recursion through massively parallel neural network and genomic
systems raises the question if it obeys the Weyl’s Law of Fractal Quantum
Eigenstates, or when derailed, pathologically results in aberrant methylation or
chromatin modulation, the root cause of cancerous growth. The growth of fractal
Purkinje neurons of the cerebellum is governed by the aperiodical discrete
quantum system of sequences of DNA bases, codons, and motifs. The full
genome is fractal; the discrete quantum system of pyknon-like elements follows
the Zipf-Mandelbrot Parabolic Fractal Distribution curve.
The Fractal Approach to Recursive Iteration has been used to identify fractal
defects causing a cerebellar disease, the Friedreich Spinocerebellar Ataxia – in this
case as runs disrupting a fractal regulatory sequence. Massive deployment starts by
an open domain collaborative definition of a standard for fractal genome dimension
in the embedding spaces of the genome-epigenome-methylome to optimally diag-
nose cancerous hologenome in the nucleotide, codon, ormotif-hyperspaces. Recur-
sion is parallelized both by open domain algorithms, and also by proprietary
FractoGene algorithms on high performance computing platforms, for genome
analytics on accelerated private hybrid clouds with PDA personal interfaces,
becoming the mainstay of clinical genomic measures prior and post-cancer inter-
vention in hospitals and serve consumers at large as Personal Genome Assistants.
Introduction
Agenda: The Cerebellum as the Platform for the Unification ofNeuroscience and Genomics by the Geometric School of Biophysics
Our understanding of both the genome and the brain will remain partial and disjointed
until we reach a unification of the intrinsic mathematics of structuro-functional
geometry of both – as the first is without question a foundation of the second.
The cerebellum emerged in the past half a century as the best known neural net of
the brain since Moruzzi (1950), Jansen and Brodal (1954), Dow and Moruzzi (1958),
and Eccles et al. (1967). Thus, this CNS subsystem became a fertile ground of
theoretical advances as recently reviewed (Manto 2008). It is remarkable that some
of the earliest concepts as shown below can be traced back to centuries, but later they
became heavily influenced not only by their underlying philosophies, but also by trendy
schools from various periods of history. It was only recently that concepts consolidated
into mathematically sophisticated theories of neural networks. For a recent review,
see Fiori (2008). References to Tensor Network Theory (TNT) are too many to list.
The level of mathematical abstraction was challenging, as the dual tensor-
representation of covariants and contravariants, while fundamental in mathematics
of generalized vector (tensor) calculus (Sylvester 1853), was not well understood in
its application for sensory vectors and motor vectors, in spite of a brilliant encap-
sulation (Anderson 1990). Here, embracing the generalization of the concept for
covariant “protein signaling” RNA (noncoding; ncRNA) versus contravariant
“executory” RNA vectors (coding; cRNA), vectors also call for cross-disciplinary
1382 A.J. Pellionisz et al.
expertise. While at the introduction of TNT, Amari initially went public with a
dubious critique (see note added in proof in Pellionisz and Llinas (1985)). However,
he reversed face soon. Amari (1991) actively uses covariant and contravariant metric
tensors and Riemannian metric tensors as a foundation of “Information Geometry.”
Yet, mathematical theories of (cerebellar) neural networks had minimal impact on
neuroscience in the twentieth century for their mathematics-aversion, prized for US
aerospace application (Pellionisz et al. 1992) and Germany (Eckmiller 1990). Decades
can be lost if paradigm-shifts are not embraced in a timely manner (Kuhn 1962). An
example is that from the encyclopedic formulation of TNT (Pellionisz 1987), it took
two decades for Roy and Llinas (2007), Llinas and Roy (2009) and Fiori (2008) to
attempt to improve on it. More interestingly, attempts were aimed at making TNT
more dynamic and also to extend internal representation of the sensorimotor geometry
to the organization of the self. Note that the experimental–theoretical collaboration
34 years earlier started with “Dynamic Single Unit Simulation of a Realistic Cerebel-
lar NetworkModel” (Pellionisz and Szentagothai 1973), cited in Pellionisz and Llinas
(1985), and the hierarchy of internal representations was also laid out (pp. 268–70).
Genomics of twenty-first century might not afford to be as luxurious to
let several decades to be wasted. The rapid rise of Genome Sequencing industry
must be matched by Genome Analytics (Lander et al. 2001; Venter et al. 2001;
Church 2005; Mardis 2006; Gibbs et al. 2007; Collins 2007; Pellionisz 2008b).
After the $3Bn “Human Genome Project” there is a general realization that “ourconcepts of genome regulation are frighteningly unsophisticated” (Venter 2010).
Indeed, instead of “gene regulation” or “genome regulation” a conceptual shift to
“multidimensional coordinated genome function” is required. It is now widely
recognized that Genome Informatics simply will not do without massive computing,
requiring algorithmic mathematical approaches to program them, and the fact that
neural function arises from neural networks that are governed by genomic and
epigenomic (hologenomic) factors. As a result, some pioneers of the field of Neural
Nets swiftlymigrated to become leaders in genome informatics (Haussler 1995), and
biologists imprinted by the General System Theory (Bertalanffy 1934) a decade ago
started to claim that “Genomics became Informatics” (Hood et al. 2002; Baliga et al.
2002). Hood and colleagues (Baliga et al. 2002) immediately recognized two types of
information in coordinate regulation “The Regulatory Network for Phototrophy
Includes at Least Two Transcription Factors”. A commonmathematical underpinning
of neuroscience and genomics emerged even before the ENCODE Project led by the
US Government concluded in the imperative that “now the community of scientistshave to re-think long-held beliefs” (Collins 2007; Pellionisz 2006; Simons and
Pellionisz 2006a). With the hindrance of old dogmas defeated in less than 3 years,
The Principle of Recursive Genome Function rapidly gained ground (Pellionisz
2008a, b, 2009a, b; Shapshak et al. 2008; Chiappelli et al. 2008; Cartieri 2009;
Perez 2010; Arneth 2010; Oller 2010; Stagnaro 2011; Stagnaro and Caramel 2011;
Elnitski et al. 2011).
With the advances of twenty-first century genomics, the cerebellum is not just
a neural net for sensorimotor coordination, but lends itself to be a unique platform
for unification, on how genomic and epigenomic (hologenomic) factors create the
61 Recursive Genome Function of the Cerebellum 1383
physiology as well as pathology of cerebellar organelles (most remarkably,
Purkinje neurons), organs (the cerebellum), and organisms (sensorimotor system).
While the geometrization of Neuroscience with TNT to arrive at the “Galileancombination of Simplification, Unification, Mathematization” (Churchland 1986)
emerged decades too early, mathematization of Genomics is now an urgent socio-
economic necessity. Without advanced mathematics yielding software-enabling
algorithms, duties of genomics are impossible to carry out within the narrow
boundaries of limited domains. This does not mean, of course, that established
disciplines are not to stay, but as Erez-Lieberman et al. (2009) present in their paper
amounting to a call by co-author Dr. Lander “Mr. President, the Genome isFractal!,” biochemistry was applied to advance, rather than hinder, a paradigm-
shift of the early seminal idea of fractal DNA folding (Grosberg et al. 1988, 1993).
TNT may qualify as the best platform for unification, from neural nets to
genomics “top-down” and “bottom-up” toward consciousness. Beyond establishing
encyclopedic use of tensor geometry (Pellionisz 1987), TNT is experimentally
supported for arm-movements by Gielen and Zuylen (1985), Bloedel et al.
(1988), and Laczko et al. (1988); for gaze control by Pellionisz (1985a), Daunicht
and Pellionisz (1987), Pellionisz and Graf (1987), and Pellionisz et al. (1991); for
vestibulo-collicular sensorimotor system by Laczko et al. (1987), Peterson et al.
(1987, 1989), and Lestienne et al. (1988). Belated followership improved upon the
pioneering (Roy and Llinas 2007; Fiori 2008).
There is not much question that growth of neural networks, such as those of the
cerebellum, is governed by genomic and epigenomic (hologenomic) factors. Like-
wise, it seems to be beyond reasonable doubt that both genome function and the
function (sensorimotor coordination) are deeply rooted in recursion (see the epoch-
making concept of “feedback” by Cybernetics (Wiener 1948)).
Our understanding of the intrinsic mathematics of both Neuroscience and Geno-
mics has reached the critical mass of mathematical overlap of these two fields of
biology. This chapter aims at an algorithmic unification of both neuroscience and
genomics by the mathematical means of non-Euclidean tensor and fractal geome-
try. HoloGenomics unites Neuroscience with Genomics, Epigenomics, in terms of
Informatics. Time has come to identify the common geometric roots of genome
function and how they govern growth and functioning neuronal networks in both
a physiological as well as a pathological manner.
Recursion in the Cerebellum
Review: Philosophies, Theories, and Computational Models asFoundations of the School of Cerebellar Recursion
Western philosophies traditionally embraced the age-old “arrow-model” of deduc-
tive, deterministic timeline and unidirectional “cause and result” (Churchland
1986; DuPre and Barnes 2008). This is in contrast to the inductive yin-yang of
equilibrium, oscillations, and interdeterminism of Eastern philosophies Zuangzi
1384 A.J. Pellionisz et al.
(�400 BC). Theory of Relativity by Einstein and the Principle of Uncertainty by
quantum mechanics of Planck-Heisenberg-Schr€odinger, therefore, shook the intel-
lectual foundations of Western philosophies.
The result was an interesting fork. On one hand “System Theory” was outlined
as an attempt to encompass complexity (Bertalanffy 1934). However, Systems
Theory hardly aimed at defining the intrinsic mathematics of living systems.
Thus, on the other hand, massive simplifications occurred, to regain temporary
balance. Compared to Schr€odinger’s “What is Life?” (1944), too early to know the
A, C, T, and G quanta of “heredity encoded by covalent bondings on an aperiodical
crystal,” Cybernetics (Greek, “to govern by feedback”), Wiener (1948) took
a “reverse engineering” trend of simplification, almost exclusively based on “feed-
back.” Cybernetics, for its reductionism and relying on concepts of engineering,
logical calculus, and information theory (McCulloch and Pitts 1943; Shannon 1948)
rose with the catapulting digital computing architectures to attain intellectual
dominance (Neumann 1958) (see also second edition with Introduction by Drs.
Churchland). Though Neumann, the inventor of computers, warned that the math-
ematics of computers and brains are profoundly different (the latter remained
a mystery with von Neumann’s demise) (1958), his tragically short-lived life
aborted the breakthrough to find mathematics intrinsic to neural (let alone the at-
that-time largely unknown genome) systems. Instead, an even more drastic simpli-
fication was dogmatized by Crick (1956/1970), groundlessly proclaiming the
DNA > RNA > PROTEINS to be “an arrow-type open loop.” Crick’s Dogma
oversimplified even the “arrow-model” into a single channel of action – clinched by
rendering recursion to “Junk DNA” pointless (Ohno 1972) – though “noncoding”
DNA is actually 98.7% of the human genome! Crick’s Central Dogma and Ohno’s
Junk DNA obsolete notions were surpassed by The Principle of Recursive Genome
Function Pellionisz (2008a, b, 2009a). Crick said (1970) that if his Central Dogma
would be proven to be untrue, it would be necessary to put genomics onto an
entirely new intellectual foundation. Now with The Principle of Recursive Genome
Function Pellionisz (2008a) not only it became demonstrably untrue but was
superseded by a more advanced theory. The revolution lies in recursion.
The sidetrack of simplification continued into the overly ambitious notion that
science does not need to understand Nature’s systems to mimic them. Thus,
“Artificial Intelligence” (AI) emerged (see the “Perceptron” Minsky and Papert
(1969)). AI took off and ruled, in part (by mathematically mistakenly) “proving”
that neural nets are incapable of performing the key exclusive “or” operation in
mathematical logic. It took pioneers of Neural Nets (Hopfield 1982) to rectify the
damage caused by the simplistic course, yet AI was only as recently as in 2003
declared by its originator ineffective.
These trends influenced the expression of the most widely accepted classical
concept (Flourens’ 1824) that the biological neural networks of the cerebellum
function to coordinate sensory and motor information. (For the general audience,
seeWikipedia and for specialists a recent review (Manto 2008)). The seminal concept
was traced back by Finger (1994, pp. 211–121) to a quarter of a Millennium to
originate implicitly the experiments by the surgeon to Napoleon, whose experiments
61 Recursive Genome Function of the Cerebellum 1385
“involved inserting needles into the brains of some pigeons . . . needles pushed to thecerebellum caused his bird to sway. . . ., this probably was one of the first experimentaldemonstrations of the association between cerebellar damage and problems withcoordination.” As shown in Finger (1994), Rolando (1908) followed-up on the
surgeon (D.J. Larrey), but largely missed, till Flourens pinned down close to
200 years ago (1824) that “I have shown that all movements persist after ablation ofthe cerebellum; they lack only being regulated and coordinated” (pp. 292 and 212 of
Finger (1994)). In modern times, Holmes (1939) re-established the concept that
without the cerebellum, coordination is known to be absent – the syndrome aptly
named, even before its metric tensor function was discovered as dysmetria.Cerebellar theories were reviewed in Pellionisz (1984); for a recent review see
D’Angelo et al. (2010). In the decades of conceptual confusion caused by the conflict
of philosophies, the already long-established facts of cerebellar coordination slipped
into “arrow models” of cerebellar theory, such as describing it as a now-known
conceptual oversimplification of a “timing device” to set the temporal distance from
intention to action (Braitenberg et al. 1967). Perhaps due to the emergence ofMinsky’s
Perceptron as a “learning device,” Marr’s model (1969) utilized a coincidence for
“motor learning.” However, he repudiated his concept, since (as he said) motor
learning did not explain coordination (Marr 1982, p. 14). With the untimely decline
of his health, a Marr-Albus “learning model” emerged (Albus 1971 see review in
Pellionisz 1986). Featuring the cerebellum as any kind of a “filter device” belongs also
with the category of “arrow models,” since it streamlines the arrow-process of
unneeded factors but disregards the cardinal notion of feedback. The cerebellum is
conceptually not a timer, not a filter, but a transformer, converting themultidimensional
vector-expression, from covariant intention tensor to coordinated contravariant
execution tensor by means of recursion (see section “Tensor Network Theory:
Vector–Matrix Recursion as Basis of the CerebellumActing as a Sensorimotor Metric
Tensor” as discussed below).
The Concept of Coordinates and Their Recursion as Basics of TensorNetwork Theory of Cerebellar Neural Nets
In a mathematical sense, as reviewed earlier, Pellionisz (1984), the geometric school
of thought about brain function, including coordination, reaches back about 400 years
to Descartes (1629). Descartes’ insert in Fig. 61.1 (ab12), from Pellionisz (1984),
illustrates his most reasonable idea – in retrospect – that by the Cartesian coordinates
both key concepts of living systems were comprised. Both multimodal compositions
of entities were shown, as well as a functional recursion of information; see the
finger-movement under the feedback of visual control. TNT “simply” generalized the
Cartesian x, y, z (and t) coordinates of the Minkowski-spacetime manifold, where it
became evident that in non-orthogonal expressions generalized vectors (tensors)
profoundly differ if expressed in a “sensory or motor manner.”
The tensorial scheme in Fig. 61.1 uses a (minimal) 2-component sensory vector
and a higher (3) dimensional motor vector that expresses the same physical object
1386 A.J. Pellionisz et al.
(invariant, in this case a displacement). The scheme illustrates the contravariant motor
efferent vector, as well as the recursing covariant proprioception vector. As explained
in detail in Pellionisz (1984) (see also Fig. 61.2 here), this recursion converges in the
brain stem in the Eigenvectors that are essential to build the cerebellar metric, as the
matrix-product of Eigenvectors, found by recursive oscillatory tremor. Even this
schematic representation points out that the interim oversimplification (e.g., exem-
plified by Lorente de No (1933) that three-neuron reflex arcs carry a one-to-one
representation) is mistaken (Szentagothai 1949). Single “loops,” for example, reflex
arcs are surpassed by a many-to-many network of neural nets, harboring some
massive interconnections, described by vector–matrix tensor geometry.
Switching to the seminal work of Genomics (Mendel 1866) regarding many
phenotypes (he investigated seven characteristic inheritable traits, in parallel),
a similar one-to-one oversimplification ensued; a decided effort to associate with,
or rather, to pin “one phenotype on a single genotype.” A key message of this
Fig. 61.1 An example of specific “System Theory” identifying the modern mathematics of
Descartes’ classic concept of “coordinates” (ab12). Surpassing the Cartesian frame of reference
by generalized coordinate systems used by Nature for sensorimotor coordination (de13), cerebellar
coordination is explained in terms of tensor geometry (Fig. 1 from Pellionisz (1984)). For
biological organelles, organisms, and organs, in this case that of cerebellar sensorimotor system,
no “Biological System Theory” will be “software enabling” unless the intrinsic mathematics is
identified, as it is shown here, or better. Further explanation is in the text, and the mathematical
procedure is elaborated in Pellionisz (1984)
61 Recursive Genome Function of the Cerebellum 1387
Fig.61.2
(continued)
1388 A.J. Pellionisz et al.
chapter is to draw a parallel that the “single gene-to-single phenotype” approach is
likewise futile, as is a “loop-type” single reflex. Instead, there is a “neural net-
work”–type “many-to-many” interactions among, say n, phenotypes, and the under-
lying, say k, genotypes. It is strongly believed that the cerebellum, with its already
modeled multicomponent factors is the best platform to sort out the underlying
mathematics of “multicomponent dual representation of covariant andcontravariant functors (defined as objects that relate categories)” of not only thecerebellar neural networks, but also of their genomic roots. In order for this to
�
Fig. 61.2 Icons of Tensor- and Fractal Geometrical Recursion in Neuroscience for a Geometrical
Unification of Neuroscience and Genomics using cerebellar cellular and network systems. For
originals and explanation in detail see neuroscience Icons (from TNT) for Icon 1 Pellionisz and
Llinas (1980), and Icons 2–3 Pellionisz and Llinas (1985). For the bottom (genomics) Icons 4–5,
for Icon 4. see Pellionisz (2008a, b), and for Icon 5 see Pellionisz (2002, 2003) and Simons and
Pellionisz (2006a). Icon 1 comprises the dual valence of vectors, if using non-Cartesian (gener-
alized, non-orthogonal) coordinate systems. The covariant and contravariant vector components
are shown in Panel A together, while in Panel B separately. It is cardinal that they are not the same
either in their values or in the way how they represent the invariant, the covariant components can
be independently measured, but they do not add physically, while contravariants are
interdependent, but they do physically generate the invariant. Panels C–F in Icon 1 show that
the covariant and contravariant vectors can be converted to one-another by themetric tensor. Icon 2:
Panel A shows that the covariant proprioception vectors could recur through the brain stem (even
without a cerebellum or sensorimotor cortex) via mossy fibers to cerebellar nuclei, and could be
directly (but inappropriately) used as if they were true contravariant executor vectors. Thus,
a recursion may take place, as shown in Icon 3. Close inspection shows that in a non-orthogonal
system of coordinates, starting with any vector (even a noise vector), after several recursion the
covariant and contravariant expressions converge into the Eigenvectors (where the incoming and
outgoing vectorial components are the same). Physically, this mechanism is an uncontrolled but
convergent tremor, mathematically the discovery of Eigenvectors. Dyadic products of Eigenvec-
tors, yielding a matrix, create the metric tensor. Icon 4 shows The Principle of the Recursive
Genome Function, that permits a DNA> RNA> PROTEIN>DNA. . . recursion (after discardingthe obsolete notions of Central Dogma and Junk DNA). Note, that for the purposes of simplicity
Icon 4 shows the recursion as a single circular line – but it symbolizes multicomponential
(vectorial) entities. The cardinally important Generalization of Recursion (from neuroscience to
genetics) is the concept introduced here that the coding DNA vectors (many “exons” acting
together), when transcribed, create RNA vectors that are of contravariant valence, since their
translation into protein vectors creates physical objects. However, when protein vectors are
signaled (measured) by noncoding DNA via bonding not only to homeodomains but also to
ncDNA vectors, they are transcribed into another RNA vector, this time of covariant (sensory)
valence. Thus, a recursion, similar to one shown in Icon 3 converges into the Eigenstates of the
recursion in the genome, and the cRNA and ncRNA Eigenvectors produce the metric, comprising
the functional geometry of the genome function. If the recursion converges to follow the Weyl’s
Law on Fractal Quantum Eigenstates, the genomic recursion switches the growth of fractal protein
structures (such as a Purkinje neuron, shown in Icon 5) into the next step of recursive hierarchy.
The physiological process requires canceling (methylating) ncDNA segments perused in the
recursion (see Fig. 61.9), such that the ncDNA fractal segments, governing growth according to
FractoGene are not overused. It follows, that hypo-methylation and incorrect chromatin modulation
could permit an uncontrolled (cancerous) growth as shown in Fig. 61.9 (yellow “cookie”).
For further details, consult the original papers containing Icons 1–5 and the text of this review,
relating the seminal concept of generalization of recursion described in Fig. 61.2 with the fractal
recursive iteration shown in Fig. 61.9
61 Recursive Genome Function of the Cerebellum 1389
happen, science needs to specify the mathematics that underlie “a biological system
theory” (Bertalanffy 1934). Identification is essential for both the neural network
and for the underlying genome, including the suggestion here that they are concep-
tually identical.
Further, as it is suggested here, science needs to move away from a “one-to-one”
and “arrow-type” mapping toward the “many-to-many more” and “recursive” anddual representations. This is important not just for theory, but for entire industries.
The “Big Pharma” model of “one gene, one disease, and one billion dollar pill” is
obsolete for over a decade because of a simplified and incorrect “one-to-one”
assumption. Now the future lies in the generalization of covariant and contravariant
neural network representations for the genome-epigenome (hologenome) system.
The cerebellar biological neural networks, as shown, provide a precedent for this
mathematical insight that is also applicable to genomics.
Generalization of Recursion from Cerebellar Neuroscience toGenomics; Covariant and Contravariant RNA Functors and TheirEigenstates
As it was shown over three decades ago in Pellionisz and Llinas (1980) if using non-
Cartesian (generalized, non-orthogonal) coordinate systems (see Icon 1 in
Fig. 61.2), invariants (such as displacement) are represented in with a dual valence.
The orthogonal projection-components, named covariant tensor-components in
mathematics by Sylvester (1853), can be independently established; however,
covariant components do not physically assemble the invariant. In turn “motor
expressions,” expressed as interdependent parallelogram-type coordinates, that he
called contravariant tensors, do assemble the object in a physical manner. It is
cardinal in mathematics of generalized coordinates (tensor geometry) that a matrix
can convert the “covariant sensory intention vectors” into “contravariant motor
execution vectors” (see Panels C-F of Icon 1). The matrix that does this is the many-
to-many interconnection-system of a massively parallel neural network of the
cerebellum. Thus, the cerebellar sensory-motor coordination is accomplished by
the conversion via the metric tensor. The metric comprises the geometry of the non-
Cartesian multidimensional space-time, embedding both sensory and motor events.
This perhaps difficult but cardinal concept of sensory- and motor components as co-
and contravariant vectors was most lucidly encapsulated by Anderson (1990)
pp. 351–355, in Anderson et al. (1990).
If dual, covariant, and contravariant functors, shown in Fig. 61.2 Panel 3, are
freely let to recur (when proprioception vectors are directly used by recursion as if
they were execution vectors, without the cerebellar cortex, see Icons 1–2 in
Fig. 61.2), they converge into the Eigenstates (where the normalized covariant-
and contravariant representations are identical – while in general they are different).
Finding the Eigenvectors characterizing the Eigenstates by free recursion (that in
sensorimotor systems is manifested in uncoordinated, oscillatory movements) is
essential, since the metric tensor (and its inverse, or Moore-Penrose Pseudo-inverse
1390 A.J. Pellionisz et al.
for overcomplete space (Pellionisz 1984)), capable of converting covariants to
contravariants, is obtained as a matrix-product of the Eigenvectors.
In category theory, covariant and contravariant aswell asmixed valence of functors
(vectors and generalized vectors, tensors, are just one specific type of functors;
they relate invariants to coordinate axes) are both well established and reaffirmed
(Francis 2008). Herein, with the conceptual guidance of Icons 3–4 in Fig. 61.2, the
dual representation is generalized to the interpretation of RNA system in Genomics.
There is no debate that the so-called amino-acid-coding RNA-s (cRNA-s, as
a multicomponent, vectorial entity) physically aggregate physical objects (proteins).
Thus, the valence of cRNA-s is contravariant, similar tomotor vector components that
also have to assemble the physical object. The contravariant cRNA vectors, however,
via RNA self-replication (Glasner et al. 2000) are available not only to construct
proteins, but to interact (interfere) with the rather different covariant multicomponent
ncRNA vectors. These measurements, what protein systems are already built, arise
when proteins bind to noncoding DNA (both in intergenic and intronic sequences)
involving transcription factors (Kornberg and Baker 1992). Through the arising
ncRNA functors (multicomponent vectors), the already built proteins are thus “mea-
sured” not just by a single sequence, referred to as “homeoproteins” generated by
a “homeodomain” (Foucher et al. 2003), but a single protein-component is signaled by
the many components of even a single but multicomponent ncRNA covariant vector.
Compare the concept to covariant sensory vectors providing independent measures of
motor events in Icon 1 of Fig. 61.3.
Putting the RNA system here into a new conceptual framework, also re-defining
the role of intronic and intergenic “noncoding” (formerly, “Junk”) DNA, recalls
earlier metaphors. Interpreting the RNA system as a “hidden layer,” an implication
referring to interconnections known in neural nets for decades Mattick (2005)
phased out his earlier metaphor that conceptually compared the RNA system to
the man-made “operating system of computers” (Mattick 2001). Recently, even
“genomic matrix” relating to fractals and chaos (Petoukhov and He 2010) and even
“RNA matrix” approaches emerged (Izzo et al. 2011). However, the co- and
contravariant valences of RNA functors have not been recognized to date. This
generalization of valence of functors from sensory- and motor vectors to covariant
as well as contravariant RNA multicomponent entities provides an opportunity to
approach the role of RNA systems in coordinated genome function in a novel
manner; that is, both conceptually and mathematically already identified in living
systems (cerebellar neuronal networks) that the genome-epigenome system is
known to generate.
Appreciation of the valences of RNA functors opens new vistas beyond
approaching the RNA-metric from a mindset that moves the perspective of science
beyond man-made technologies like operating systems of computers (Mattick
2001). Looking at the RNA system in a new light as “the metric tensor of protein
building genic sequences regulated by protein sensing noncoding sequences,” the
RNA system is conceptually likened to a “genomic cerebellum.” First of all, this
permits deploying already proven advanced geometric (thus software-enabling)
analysis of experimental results of genomics. Second, the perspective on evolution
61 Recursive Genome Function of the Cerebellum 1391
is affected by recalling the shifting metaphor by Mattick from “operating system”
to “hidden layer” (2001 vs 2005) and his reference that the RNA system serves
a “coordinated genome expression.” One cannot help noticing that the “invention”
through evolution of the physically separate, additional cerebellar neural network
(with the shark) provided for a new class of more highly coordinated vertebrates.
The conceptual equivalence is noted, therefore, that much of single-cell organisms
Transformation of P[xi,yi,(zi)]
XiYiZi
X
y
x
P3
x3x2
x1
P2�
P1�
P3�
P2P1
V2 u2 u3
u1
x
y
θ1
θ2θ3
V1
V3
Y
S �
Z
xiyi1
wi •
explicitly fractaltiling
implicitly fractalcoherence explicitly fractal
network
(u1,u2,u3) = (v1,v2,v3)•
•=xiyi1
xi�
yi�
1=
aici0
bidi0
eifi1
xiyi1
wi� • •=
a00a10a20
a01a11a21
a02a12a22
xiyi1
rxiryi1
xi�
yi�
1=
cosθsinθ
0
−sinθcosθ
0
Fig. 61.3 The conceptual linchpin of multidimensional generalized vector–matrix approach
(d) to the Fractal Approach (a, b, c) (From Bieberich 1999)
1392 A.J. Pellionisz et al.
contain a minimal amount of “noncoding DNA” – thus appear to operate with
minimal covariant ncRNA, similar to organisms before the cerebellum appeared,
permitting only an imprecise, un-coordinated execution of genomic commands.
As the amount of noncoding (regulatory) sequences hyperescalated, the emerging
RNA-metric permitted the coordinated growth and governance of complex (also
multicellular) organisms. This new interpretation of the RNA system is to be
compared to Mattick’s referral to “the Cambrian explosion” (Mattick 2004).
At the least, identification of a common advanced geometry intrinsic to living
systems makes “System Theory” approaches to genomic systems mathematically
explicit. A more remote but an inevitable goal for the use of a common advanced
geometry is to accelerate the unification of genomics and neuroscience. It is fully
realized that building this seminal idea into a robust school of thought will require
significant time and resources.
Recursive Algorithms Rule Both Vector–Matrix and FractalRepresentations
Algorithms based on recursion (see Icon 4 of Fig. 61.2 from Pellionisz (2008a, b))
share the fundamental property that each state of the system is deduced from its
previous states. Recursion, in itself, does not discriminate analog (e.g., traditional
feedback) mechanisms from digital deduction as, for example, in the sequence of
Fibonacci numbers, where each subsequent integer is the sum of the previous two.
The metric tensor characterizes the non-Euclidean geometry with integer dimen-
sions, established by recursion of covariants to contravariants to compose the
metric from Eigendyads (Pellionisz and Llinas 1985). The embedding
Minkowski-spacetime manifold, however is “smooth,” mathematically speaking
it is derivable. However, Purkinje neurons show a non-Euclidean, moreover,
a discrete geometry with fractal (non-integer) dimension (Pellionisz 1989).
Realization that the same cerebellum utilizes recursion of dual vectors, as well as
its main type of neurons, the Purkinje cells are built by an also recursive, but by
a rather different fractal iterative recursion (see Icon 5 of Fig. 61.2; Pellionisz 2002,
2003) a cardinal question arose ever since the fractal model of Purkinje neurons
(Pellionisz 1989). The question became even more vexing with the FractoGene
concept, stating that fractal DNA governs growth of fractal organelles such as the
Purkinje neuron; fractal organs such as the lung, circulatory systems; and organisms
such as the Cauliflower Romanesca pictured in Pellionisz (2008a). The question
was if the vector–matrix and fractal representations are in a mathematical conflict
with one-another, or rather, if they reveal another profound dualism, similar to one
already encountered in physics.
The question was also conceptual regarding not only the mathematics, but also
possibly referring to a “language.” The “early wave” of looking at fractality of DNA
suspected it as a “language” (Flam 1994). The concept of a “language,” however,
does not appear to be consistent with the concept of “sensorimotor coordination.”
Resolution of the question became easier once the “hint” that fractality reflects
61 Recursive Genome Function of the Cerebellum 1393
a “language”was dismissed (Chatzidimitriou-Dreismann et al. 1996). Section on “The
Genome is Fractal! Proof of Concept and the Basis of Generalization:Whole Genome
Analysis Reveals Repetitive Motifs Conforming to the Zipf-Mandelbrot Parabolic
Fractal Distribution Law of the Frequency/Ranking Diagram” shows below that the
established fractality of the genome conceptually supports FractoGene, fractal growth
of Purkinje cells governed by fractal DNA. Both in the DNA and in networks of
neurons, the fractality characterizes the geometry in a consistent manner.
The question was settled by Bieberich (1999) (see his Figure reproduced
as Fig. 61.3 in this chapter) to show a conceptual consistency of fractal and vector–
matrix representations. Thus, a geometric characterization of sensorimotor function
and the geometry of the Purkinje neurons that implement smooth (derivable) function
by non-derivable fractals are not only compatible, but mutually convertible. The
revelation by Bieberich (1999) was not entirely surprising, given the known fact in
physics that light can be seen as a wave-phenomenon, or particle-phenomenon,
depending on the theory of Schr€odinger or Heisenberg. Thus, the Bieberich-diagramis intellectually rather pleasing. Even more intriguing is its extension toward fractal
internal representation (consciousness) in Bieberich (2011).
Based on insights from fractal modeling of Purkinje neuron (Pellionisz 1989),
utilities could be developed based on the of fractality of both DNA and the
organelles, organs and organisms grown by the genome, the concept of FractoGene
by Pellionisz (2002, 2003, 2006). The FractoGene algorithmic approach to the
whole genome provided quantitative predictions that could be verified or refuted by
experimentation; moreover the “Fugu Prediction of FractoGene” (that the 1/8 of
the noncoding DNA of fugu compared to that of the human should result in
a “fractal primitive” dendritic tree in the fugu) was supported by experimental
results (Simons and Pellionisz 2006a, b).
Tensor Network Theory: Vector–Matrix Recursion as Basis of theCerebellum Acting as a Sensorimotor Metric Tensor
Recursion of sensory to motor vectors (and the generalization of valence of RNA
functors) was characterized by Icon 2 of Fig. 61.2 as an essential procedure to
converge into Eigenvectors, with their matrix-product comprising the geometry in
the metric tensor. With the example of encyclopedic Fig. 61.4 of this chapter from
Pellionisz (1987), it is shown how suchmetric is the basis of an entire system of gaze
control, stabilizing the head by the vestibulocollic sensorimotor neural network.
Icons 2–3 of Fig. 61.2 showed that sensory functors could recur directly, used in
an unchanged manner, as motor functors. However, the recursion would result in an
oscillation converging into Eigenvectors. In the cerebellar sensorimotor system,
the Eigenvectors are imprinted in the inferior olive (Pellionisz and Llinas 1985). In
turn, as shown in Fig. 61.4 here, Eigenvectors from the inferior olive give rise to
their matrix-product implemented by the neuronal network of cerebellar cortex.
The scheme shown in Fig. 61.4 stabilizes gaze (head position) by a two-step
operation: First, there is a covariant embedding from a symbolically
1394 A.J. Pellionisz et al.
two-dimensional sensory vector into an also covariant, but higher (figuratively, 3)
dimensional motor intention vector (i) – that would go directly to (mis)serve as an
imprecise execution vector (since motor vectors must be contravariant; (i) should
be (e)). Through the ascending mossy fibers, the (i) covariant intention vector is
both converted into the (�e) contravariant vector (negative, since Purkinje cells are
inhibitory), that with the mossy fiber collateral (i) vector in the cerebellar nuclei
constitutes an output vector (i�e). Thus, the brain stem would send out instead of
the covariant intention vector (i) the proper e ¼ i�(i�e) precise contravariant
execution vector. This architecture explains why the entire sensorimotor would
work (as for a dysmetric patient; even Purkinje cell affected only by alcohol) with
intentions directly executed, but the additional neural network that was a nifty
improvement as an addition to the brain of the shark makes a dysmetric direct
execution of intentions into one that matches the physical geometry of the executor
system (in this case, muscles) with its internal metrical representation.
Fig. 61.4 Tensor network model of the vestibulocollic reflex, embodying a covariant intention to
contravariant motor execution transformation via the cerebellar neuronal network (From
Pellionisz (1987)). For details, see the original publication and the text below. This figure also
serves as the inspiration of the seminal concept of generalization of TNT to Genomics. The
generalization is based on the fact that a physical object of the head movement is both measured
by the covariant sensory vector that converted both in dimensionality and covariant to
contravariant valence. Likewise, the genome expresses physical objects (proteins) both by pro-
tein-coding codons (in a contravariant manner), that can be measured by similar (but noncoding
triplets, wherein the detection is covariant), but in order to attain quantum fractal eigenstates of
stable protein systems a many-to-many RNA converter is needed. The RNA system is, therefore,
conceptually equivalent to the sensory-motor transformer of the cerebellum
61 Recursive Genome Function of the Cerebellum 1395
Coordination of an entire sensorimotor architecture is presented here to illumi-
nate how non-trivial the generalization of contra- and covariant cRNA functors
directly recursing into ncRNA covariant functors is. Such direct recursion is
excellent for finding the Eigenstates of a DNA > RNA > PROTEIN recursive
system, but the multicomponent RNA Eigenvectors must interact in an all-to-all
manner, bymeans of RNA interference, not just of one component, Fire et al. (1998),
but in a many-to-many multicomponent manner. Also, the sensorimotor coordina-
tion scheme is to illuminate why RNA interference is “silencing” – conceptually
similar to the inhibitory effect of cerebellar Purkinje cells.
Development of the school of functional geometry of a comprehensive system of
coordinated genome function, comparably to that of a sensorimotor apparatus,
requires a long-term program. One of the most difficult questions is if the genomic
recursion obeys the Fractal Weyl’s Law on Fractal Quantum Eigenstates (see
Shepelyansky (2008), originally Weyl (1912)). This question will be discussed in
the section on “Future Directions.”
Fractals Are Pervasive in Nature; Both the Cerebellar Brain Cells andthe DNA are Fractal Objects
Mandelbrot (1983) coined the term “fractal” in his epoch-making book only about
a quarter of a century ago, but the impact of identifying fractal geometry intrinsic to
Nature is already profound.
The Zipf-School Suspected that the DNA Contained a Fractal LanguageThe first “hints” that the A, C, T, and G nucleotide sequences of DNA (especially
of noncoding DNA) possibly harbored a (mathematical) “language” was published
before the epoch of “massive whole genome sequencing,” in 1994 in Science
(see Fig. 61.1 in Flam 1994). Its original caption: “Line of evidence. Plottingfrequency against rank of arbitrary ‘words’ in noncoding yeast DNA yields thelinear plot found in human language” reveals the key word “arbitrary.” Note that
“words” of the noncoding DNA were three to eight bases, sampled in an unjustified
manner. Neither graph appeared to conform to the straight “Line of evidence” of
Zipf’s law.
The study reported by Flam was based on a comparison with the empirical
“Zipf’s law,” which applies to natural languages Zipf (1949). The distribution of
frequencies (actual occurrences) of words in a large corpus of data versus their rank
is generally a power-law distribution, with exponent close to one. Zipf’s law is thus
an experimental law, not a theoretical one. Zipf-like distributions are commonly
observed, in many kinds of phenomena. However, the causes of Zipf-like distribu-
tions in real life are a matter of some controversy, with DNA being no exception.
While the early observations applied to DNA in 1994 were found worthy of
reporting in Science and were widely heralded that “something interesting was
lurking in the junk (DNA),” the “Zipf-test” was inconclusive. Review by Simons
and Pellionisz (2006a) pointed out that investigators failed to detect “well-defined
1396 A.J. Pellionisz et al.
scaling or fractal exponents” (Chatzidimitriou et al. 1996) or “any signs of hidden
language in noncoding DNA” (Bonhoeffer et al. 1997).
Empirical law aside, the biggest problem was the definition of “words” in the
DNA. First, Harvard linguistics professor Zipf (1902–1950) established his “law,”
based on observations on the English language, in which “words” are taken for
granted. He found that in text samples the frequency of any word was roughly
inversely proportional when plotted against the rank of how common each word
was; the frequency of the k-th most common word in a text was roughly propor-
tional to 1/k. Plotting both frequency and rank on a logarithmic scale, “Zipf’s law”
was expected to yield a declining linear graph also for “words” of the DNA.
When applying this natural language linguistics to DNA the results were not
entirely convincing (Fig. 61.1 of Flam 1994). The problem was not only that the
graphs did not quite conform to the linear Zipf’s law. IIt tsaasdIt is unacceptable
that the definition in the noncoding DNA was completely and explicitly arbitrary.Of course, there was no definition at that time of what A, C, T, and G strings might
constitute “words.” In the analysis conducted by Mantegna et al. (1994): “when thegroup arbitrarily divided up their samples of junk (DNA) into “words” between3 and 8 bases long and applied the Zipf test, the telltale linear plot emerged.”
Looking at the reproduced Fig. 61.1 of Flam (1994), the plot (for noncoding
DNA “words” open squares on a log-log scale) starts fairly close to linear, but drops
off remarkably at the tail end. The original Flam diagram of the Zipf’s law for DNA
was even more controversial when it was applied to the “coding regions” of the
DNA (see graph of open circles in Fig. 61.1 from Flam 1994). Here, Flam claimed
that the Zipf’s law “failed” – and the reason cited was that “The coding part (of theDNA) has no grammar – each triplet of bases corresponds to an amino acid ina protein. There’s no higher structure to it.”
Today, both the “definition” of arbitrarily picked three to eight letter strings for
“words” and the “axiom” that there is no higher structure to coding DNA appear
demonstrably dogmatic.
Zipf’s law is most easily observed by scatterplotting the data, with the axes being
log(rank order) and log(frequency). The simplest case of Zipf’s law is a “1/f function.”Given a set of Zipf-like distributed frequencies, sorted from most common to least
common, the second most common frequency will occur 1/2 as often as the first.
The nth most common frequency will occur 1/n as often as the first. However, this
cannot hold precisely true, because itemsmust occur an integer number of times: there
cannot be 2.5 occurrences of a word. Nevertheless, over fairly wide ranges, and
to a fairly good approximation, many natural phenomena obey Zipf’s Law.
The Genome is Fractal: Grosberg-School Suspected that the DNAShowed Fractal FoldingThe classic book of the mathematician who coined the word “fractal” (as a measure
of dimension of roughness of results of recursive procedures), by Mandelbrot
(1983) generated a huge impetus into the direction of pulling away from looking
at the genome as a language, and looking at fractals more as the “geometry of
nature.” The twin schools of thought, toward approaching the structure of the
61 Recursive Genome Function of the Cerebellum 1397
genome – and the protein structures whose development it governs, manifested in
the seminal work by Grosberg et al. (1988, 1993) to claim that the folding of DNA
strands were fractal. Decades later, as an eminent example how established
methods of biochemistry can be used to support paradigm-shifts, the Science
cover article appeared (Erez-Lieberman et al. 2009), in effect the Science Adviser
to the US President, Eric Lander appealing “Mr. President, the Genome is Frac-tal!” Inspired by the Hilbert-curve, a recursive folding that provides the much
needed propensities. First, it is knot-free to permit uninterrupted transcription.
Second, it is ultra-dense to enable squeezing the 2-m-long DNA strand into the
nucleus of a cell with 6 mm diameter. Remarkably, the Hilbert-curve is capable of
filling the entire space available, and in its 3D form its fractal dimension is 3.0.
Third, it also provides the advantage that is paramount for The Principle of
Recursive Genome Function, Pellionisz (2008a, b) that the DNA can be read not
only serially, from one end to the thread to the other, but because all segments of the
DNA are in maximal proximity to one-other, they can also be read in parallel.
The Perez-School Shows that the DNA is Fractal at DNA, Codon- and FullChromosome Set and Whole Genome LevelsThe Perez-school of study of recursive systems was interdisciplinary (Perez 2011b)
and showed first results in 1988 (Perez 1988a, 1991).The fractal nature of A, T, C,
and G coding or noncoding nucleotide sequences, chromosomes and genomes was
evidenced over two decades (see review Perez 2011a). Details, for example, Perez
(1991) and Marcer (1992) are comprised in two books (Perez 1997, 2009a). The
results spanning from recursive studies through DNA and full genome analysis,
including full set of chromosome levels, Perez (2008) is likely to be a serious
candidate to the measure of “Abstract DNA Roughness” as proposed in section
“Public Domain Agenda in Industrialization of Genomics: Local and Global Fractal
Dimension as a Standard Definition for Optimally Distinguishing Cancerous and
Control Genomes Based on Their Abstract Measure of “Roughness”.”
Fractals to DNA Numerical Decoding: Toward the Golden Ratio“Small is beautiful.” Inspired by the recursive “GameofLife” (Gardner 1970) using the
largest computers in the time a cellular automata a large random 0/1 cell populations
was run in 1988 (Perez 1988a and 2009b). After 110 parallel network iterations, with
a recursive single-line code, a “clown” pattern (see Panel 1 of Fig. 61.5) emerged from
the small seven cells “U” (see upper left corner of Panel 1 of Fig. 61.5 from Perez
(1988a)). A strong illustration of « small is beautiful » is the discovery of a predictive
formula of the Mendeleev’s Elements periodic table architecture (Perez 2009a, b).
The “Fractal Chaos” Artificial Neural NetworkIn the 1980s, various parallel artificial neural networks were explored (Perez 1988a
1988b), with a particular interest in discrete waves and by fractals. The fractal chaos
is summarized by right-bottom Panel 5 of Fig. 61.5. In the dynamics of the fractal,
a curious focal point seems to emerge: the “Golden ratio.” The fractal network also
provides “deja vu” recall memory and holographic-like memory (Perez 1990a;
1398 A.J. Pellionisz et al.
Perez and Bertille 1990). At that time chaos in the DNA was also searched, but it is
discrete; A, T, C, and G bases could be coded by integers, while chaos theory is
based on real numbers. Note that the ratio between 2 Fibonacci integers is near to
the Golden ratio. This raised the question of an integer-based chaos theory. Indeed,
a hypersensitivity of the fractal for inputs based on recursive Fibonacci numbers
was demonstrated (Perez 1990b).
Fig. 61.5 Examples from the Perez-School of Recursive Results. Panel 1: “Clown” emerging
from U (upper left corner), citing original recursions in 1988 by Perez (Reproduced from Perez
(2008a)). Panel 2: DNA supracode and recursive Fibonacci series: 1 1 2 3 5 8 13 21 34 55 89 . . .Example of resonances in HUMC1A1 gene (Reproduced from Perez (2011a)). Panel 3: Chromo-
some 1–8. The Evidence of Binary Proteomics Code (red) and Modulated Proteomics Code (blue)at the Whole Human Genome Scale. Green: Genomic, Red: Proteonomic (Reproduced from Perez
(2011a)). Panel 3: Chromosome 9-Y (Reproduced from Perez (2011a)) Panel 5: Perez (2010) n.
Fractals to DNA numerical decoding: the Golden ratio. Evidence of Golden ratio hypersensitivity
in a specific region of the “Fractal Chaos” recursive neural network model (From original figure
from (Perez 1997), reproduced on Web (Perez 2009b))
61 Recursive Genome Function of the Cerebellum 1399
“DNA SUPRACODE” OverviewA connection between DNA coding regions sequences as gene sequences A, T, C,
and G patterned proportions and Golden ratio–based Fibonacci/Lucas integer num-
bers were proposed (Perez 1991; Marcer 1992, see also Fig. 61.5. Panel 2). Corre-
lation samples were established in genes or gene-rich small genomes with evolution
or pathogenicity (example of HIV genome particularly; see the book Perez (1997)).
“Resonances” were analyzed, where a resonance is a Fibonacci number of contig-
uous A, T, C, and G nucleotides (i.e., 144). If this sub-sequence contains exactly 55
bases T and 89 bases C, A, or G, this set was called a “resonance.” Thousands of
resonances were discovered (see upper right corner of Panel 2 of Fig. 61.5 from
Perez (1991)): in HIV – the whole genome is long of about 9,000 bases – there are
resonances overlapping about two third of the genome.
In Single-Stranded DNA Human Genome, Codons Population are Fine-Tunedin Golden Ratio ProportionsA new Bioinformatics bridge between Genomics and Mathematics emerged
(Perez 2010). This “Universal Fractal Genome Code Law” states that the frequency
of each of the 64 codons across the entire human genome is controlled by the
codon’s position in the Universal Genetic Code table. The frequency of distribution
of the 64 codons (codon usage) within single-stranded DNA sequences was
analyzed. Concatenating 24 Human chromosomes, it was demonstrated that the
entire human genome employs the well-known universal genetic code table as
a macro-structural model.
The position of each codon within this table precisely dictates its population. So,
the Universal Genetic Code Table not only maps codons to amino acids, but also
serves as a global checksum matrix. Frequencies of the 64 codons in the whole
human genome scale are a self-similar fractal expansion of the universal genetic
code. The original genetic code kernel governs not only the micro-scale but the
macro-scale as well. Particularly, the six folding steps of codon populations
modeled by the binary divisions of the “Dragon fractal paper folding curve” show
evidence of two attractors. The numerical relationship between the attractors is
derived from the Golden ratio. It was demonstrated that:
1. The whole Human Genome Structure uses the Universal Genetic Code Table as
a tuning model. It predetermines global codons proportions and populations. The
Universal Genetic Code Table governs both micro- and macro-behavior of the
genome.
2. The Chargaff’s second rule from the domain of single A, T, C, and G nucleotides
was extended to the larger domain of codon triplets.
3. Codon frequencies in the human genome were found to be clustered around two
fractal-like attractors, strongly linked to the Golden ratio 1.618 (Perez 2010).
A Strange Meta-Architecture Organizes Our 24 Human ChromosomesA curious interaction network was found among our 24 human chromosomes
(Perez 2011a) (see Fig. 61.5, Panels 3–4 for human Chromosomes 1–8 and 9-Y,
1400 A.J. Pellionisz et al.
respectively). It was proven that the entire human genome codon population is
fine-tuned around the “Golden ratio” (Perez 2010). Across the entire human
genome, there appears to be an overall balance in the whole single-stranded DNA.
This digital balance fits neatly around two attractors with predominant values of 1
and (3-Phi)/2, where Phi is the Golden ratio. Yet, the same analysis applied individ-
ually to each of the 24 chromosomes of humans and to each of the 25 chromosomes
of the chimpanzee which reveals a 99.99% correlation between both genomes but
diversity and heterogeneity particularly in the case of our chromosomes 16 17 19 20
and 22 (see the book “Codex Biogenesis,” Perez (2009a)). Thus, a paradox emerges.
The same analysis shows a global unity across the genome, whereas, applied to
each of the constituent chromosomes of this same genome a great heterogeneity
between these chromosomes is revealed. With the objective to analyze this paradox
in greater depth, a meta-structure was discovered that overlaps all 24 human
chromosomes. It is based on a set of strong numerical constraints based particularly
on Pi, Phi and integer numbers such as 2, 3, etc. A functionality of this fine-tuned
structure appears: the structure is 90% correlated with the density of genes per
chromosome from the Human Genome project. It is 89% correlated with the
chromosome’s permeability to intrusion by retroviruses like HIV, 94% with CpG
density, and 62% with SNP inserts/deletes. Finally, a classification network of the 24
human chromosomes was discovered, including one measuring scale, ranging from
1/Phi (chromosome 4) to 1/Phi + 1/Pi (chromosome 19), which is both correlated
with the increasing density of genes and permeability to the insertion of external
viruses or vaccines.
Unifying All Biological Components of Life: DNA, RNA, ProteinsA powerful basic Pi, Phi based numerical projection law of the C O N H S P bio-
atoms average atomic weights were established (Perez 2009a), and methods will be
published in a forthcoming paper (Perez 2012). An integer-based code unifies the
three worlds of genetic information: DNA, RNA, and Protein-aggregating amino
acids. Correlating, synchronizing, and matching Genomics/Proteomics global pat-
terned images in all coding/noncoding DNA sequences, all biologic data is unified
from bio-atoms to genes, proteins, and genomes. This code applies to the whole
sequence of human genome, and produces generalized discrete waveforms. In the
case of the whole double-stranded human genome DNA, the mappings of these
waves fully correlate with the well-known Karyotype alternate dark/gray/light
bands. This “unification of all biological components” is illustrated in Panels 3–4
of Fig. 61.5 (Perez 1988a). A complete proof of self-similarity within the whole
human genome is provided by Perez (2008). In this “binary code” which emerges
from whole human DNA, the ratio between both bistable states is exactly equal to
“2” (the space between two successive octaves in music). As shown in Perez (2008)
the Top State is exactly matching with a Golden ratio, the Bottom State is also
related to the Golden ratio. If PHI ¼ 1.618, it is the Golden ratio, and is
phi ¼ 0.618 ¼ 1/PHI, then the “Top” level ¼ phi ¼ 1/PHI and the “Bottom”
level ¼ phi/2 ¼ 1/2 PHI. Top/Bottom ¼ 2.
61 Recursive Genome Function of the Cerebellum 1401
Neural Net Elements are Fractal: Purkinje Neuron Fractal ModelAbout the same time as the Grosberg-school of thought devoted itself to the
analysis of fractal folding of DNA, the School of Recursive Function developed
a fractal structural model of a dendritic arborization (Pellionisz 1989). The seminal
concept of “recursion” to the DNA to build a fractal neuron is explicitly argued in
point 3.1.3 of that paper: “Neural Growth: Structural Manifestation of RepeatedAccess to Genetic Code”: “One of the most basic, but in all likelihood ratherremote, implication of the emerging fractal neural modeling is that it corroboratesa spatial ‘code-repetition’ of the growth process with the repetitive access togenetic code. This conceptual link between the two meta-geometries of doublehelix and ‘fractal seed’ may ultimately lead to precisely pinpointing those exactdifferences in the ‘genetic’ code that lead to a differentiation to Purkinje-, pyra-midal cell, Golgi-cell or other type of specific neurons. It must be emphasized,however, that establishing a rigorous relation of these ‘code sequences’ to thegenetic code that underlies the morphogenesis of differentiated neurons may be farin the future.”
The Genome is Fractal! Proof of Concept and the Basis ofGeneralization: Whole Genome Analysis Reveals Repetitive MotifsConforming to the Zipf-Mandelbrot Parabolic Fractal Distribution Lawof the Frequency/Ranking DiagramThis chapter decidedly expands on this point to provide support to the generaliza-
tion, to further detailing a study heralded earlier on the fractality of a whole DNA
(Pellionisz 2006; Simons and Pellionisz 2006b; Pellionisz 2009a).
With a rapidly increasing number of species in which the whole genome is
sequenced and DNA is fully available moreover “motif discovery methods” are
increasingly available. See the TEIRESIAS algorithm by Rigoutsos and Floratos
(1998), the MEME and MAST algorithms by Bailey and Gribskov (1998), and
GEMODA algorithm by Jensen et al. (2006), and Kyle et al. (2006), repetitive
“motifs” lend themselves as natural units serving as “words.” This raises not only
the necessity, but a possibility to revisit the original Zipf’s law analysis (Flam 1994;
Mantegna et al. 1994).
In the study reported here, the recently found short, repetitive sequences
(“Pyknon”-s) described by Rigoutsos et al. (2006) are used as more natural
“words” than completely arbitrarily picked three to eight nucleotide sequences.
In the human DNA, they found about 128,000 short, repetitive sequence elements,
apparently indiscriminately distributed over coding as well as noncoding regions of
the DNA.
Therefore, there is no need, indeed no basis to separate “words” occurring in the
DNA either in the regions of “genes” or what used to be called “junk” DNA. In
addition, the short, repetitive sequence motifs mined by the TEIRESIAS algorithm
Rigoutsos and Floratos (1998) showed no apparent difference in occurrence either
in the “coding” or “noncoding” regions.
Using the web-interface by the Group of Rigoutsos at IBM Watson Research
Center http://cbcsrv.watson.ibm.com/Tspd.html a “pyknon-type” motif discovery
1402 A.J. Pellionisz et al.
was made for the whole genome of the Mycoplasma genitalium, the smallest DNA
known Fraser et al. (1995).
The web-interface returned the list of short, repetitive DNA sequences in the
order of their ranking (as integers) with the frequency of occurrence (also as
integers). Results immediately lend themselves to a log-log plotting of the
frequency (y) against ranking (x), as seen in the graph below.
Figure 61.6 shows the frequency (y) plotted against ranking (x) of “Pyknon-
Like-Elements” short repetitive sequences (PLE-s) of the whole DNA of
Mycoplasma genitalium. Results reveal a “Zipf-Mandelbrot Parabolic Fractal Dis-
tribution.” Both frequencies and occurrences are shown on a log-log scale. Note
that the actual distribution is distinctly different from the linear Zipf’s law. More
detailed analysis of the above results reveals by standard curve-fitting that the data
can be modeled by the generalization of Zipf’s Law, defined as the Zipf-Mandelbrot
Parabolic Fractal Distribution. The Zipf-Mandelbrot function is given by
f ðk; N; q; sÞ ¼ 1=ðk þ qÞsHN;q;s
where HN,q,s is given by
HN;q;s ¼XN
i¼1
1
ðiþ qÞs:
This may be thought of as a generalization of a harmonic number. In the limit as Napproaches infinity, this becomes the Hurwitz zeta function z(q,s). For finite N and
q ¼ 0 the Zipf-Mandelbrot law becomes Zipf’s law. For infinite N and q ¼ 0 it
becomes a Zeta distribution.
42 3 511
100
10
0
Zipf - Mandelbrot “Parabolic FractalDistribution”
of short repetitive DNA sequences in thewhole genome of Mycoplasma Genitalium
Fig. 61.6 Zipf-Mandelbrot
Parabolic Fractal Distribution
Curve of short repetitive DNA
sequences in the whole
genome of Mycoplasmagenitalium. Frequency as
a function of rank is parabolic
on a log-log scale, after
Pellionisz (2009a). See
detailed explanation the
reference and in the text
below
61 Recursive Genome Function of the Cerebellum 1403
In the Parabolic Fractal Distribution, the logarithm of the frequency or size
of entities in a population is a quadratic polynomial of the logarithm of the
rank; standard curve-fitting approximates the data with the quadratic polynomial
y ¼ �0.052x2 � 0.0015x + 1.71.
As in typical cases, there is a so-called King effect where the highest-ranked
item(s) tend(s) to exhibit a significantly greater frequency or size than the model
predicts on the basis of the other items.
Data by the Rigoutsos et al. (2006) motif discovery reveal the Zipf-Mandelbrot
Parabolic Fractal Distribution curve of frequency against ranking of short repetitive
sequences in the entire genome (full DNA) of a free-living organism. It is note-
worthy that for the analysis no distinction between the “protein-coding” and “non-
protein-coding” DNA segments need to be made.
Nonetheless, one might argue that since the DNA of theMycoplasma genitaliumcontains only <8% “noncoding DNA” (the rest is almost a “wall-to-wall” protein-
coding sequence), the found Parabolic Fractal Distribution might be characteristic
for the coding DNA.
Therefore, utilizing recent results of identification of “FractoGem”-s (group of
“FractoSet”-s, each composed of pyknon-type repetitive short sequences that are
found strictly in the noncoding intronic regions of the Presenilin gene of
Alzheimer’s, data from http://www.fractogem.com), a comparative graph is pro-
vided below, applying strictly for noncoding short repetitive sequences (Fig. 61.8).
While the number of data-points are limited since the FractoGem of the human
Presenilin intronic areas contains 27 FractoSet formations only and each with
a maximal number of 13 “pyknon-type” short repetitive sequences, the Zipf-
Mandelbrot Parabolic Fractal Distribution curve appears applicable. Results of this
chapter may need to be reproduced and extended to the whole genome of species in
3210 4 5
1
0
0.2
0.4
0.6
0.8
1.2
1.4
1.6
1.8
2
Fig. 61.7 Curve-fitting (in purple) of the frequency (y) against ranking (x) of “Pyknon-type”
short repetitive sequences of the whole DNA of Mycoplasma genitalium (in blue). The curve
reveals a Zipf-Mandelbrot “Parabolic Fractal Distribution” that can be approximated by the
quadratic polynomial of y ¼ �0.052x2 � 0.0015x + 1.71, after Pellionisz (2009b); see detailed
explanation in the reference and in the text below
1404 A.J. Pellionisz et al.
comparative genomics to DNA other than that of Mycoplasma genitalium. Otherpoints of interest are if some coding DNA corpus larger than that of the intronic
sequence of a single gene (Presenilin) will yield similar indiscrimination for the
Parabolic Fractal Distribution, and if “motif discovery algorithms” other than
Tereisias by Rigoutsos and Floratos (1998) confirm the present study.
Overall, if a “mathematical language” is suspected to be hidden in the DNA (in
coding as well as noncoding regions), the thesis of this chapter is that currently best
candidates for “words” are the short, repetitive segments as revealed by the
Tereisias motif discovery algorithm, and the most likely mathematical language
is modeled by the Zipf-Mandelbrot Parabolic Fractal Distribution curve of fre-
quency over ranking, a log-log scale.
Conclusions
Main conclusions of this chapter are:
• The One-to-One Arrow-model of “three-neuron-reflex arc” by Lorente de No
(1933) lost to the “All-to-All” matrix-model (Szentagothai 1949) (“Elementaryreflex arcs are convenient abstractions rather than real functional units of thenervous system”).
• The Arrow-model of Genomics (Crick’s Central Dogma (1956/1970)) was obsolete
before its birth by the importance of “feedback” byCybernetics, (Wiener 1948), and
is superseded by The Principle of Recursive Genome Function (Pellionisz 2008a).
• The massively parallel systems of Neural Nets and Recursive Genome Function
are to be mathematically described by multicomponent entities (vectors includ-
ing dual representation), rather than by serial loops.
0.8 1 1.20.6
0.6
0.8
1
1.2
1.4
0.4
0.4
0.2
0.2
00
Fig. 61.8 Zipf-Mandelbrot “Parabolic Fractal Distribution” in the strictly noncoding DNA.
Frequency (y) against ranking (x) data-points are in blue for repetitive short sequences of the
FractoGem of the Presenilin intronic areas in Alzheimer’s. The curve dotted in purple reveals
a Zipf-Mandelbrot “Parabolic Fractal Distribution” that can be approximated by the quadratic
polynomial of y ¼ �0.85x2 � 0.022x + 1.32, after Pellionisz (2009b); see detailed explanation
there and in the text below
61 Recursive Genome Function of the Cerebellum 1405
• Biological System Theory is compelled to identify the mathematics of the
system, in a manner to conclude in software enabling algorithms.
• Coordination by the cerebellum is to be characterized by generalized coordinates
as in non-Euclidean tensor and fractal geometry.
Neuronal and Genomic Systems are Governed by RecursiveAlgorithms of Massively Parallel Networks, Not Only Including, butSurpassing Serial Feedback
The above main conclusions are comprised into the single above statement. The
consequences are the following:
• The cerebellar Purkinje Neuron is fractal, similarly the folding of the DNA is
fractal.
• The Zipf-Mandelbrot Parabolic Fractal Distribution curve of the full DNA of an
organism clinches that the Genome is Fractal.
• Computational Unification is made possible by the full utilization of recursion,
deploying Neural Net Algorithms.
• Neural Nets are applicable because the Recursive Genome Function is massively
parallel.
Application of Fractal Genomics is Already Here
While to most people “fractals” are either pretty pictures or some exotic branch
of mathematics, as usual in the history of mathematics, practical applications
already exist.
Friedreich Spinocerebellar AtaxiaSince the function of the cerebellum is sensorimotor coordination (by acting as
a metric tensor), symptoms of aberrant cerebellar function is often called
“dysmetria” (literally meaning that the precise metric is absent). Research of the
great number of varieties of “ataxia” such as the lack of proper cerebellar coordi-
nation is a very large, active field, as reviewed recently by Manto and Marmolino
(2009).
A specific kind of dysmetric cerebellar disorder is the Friedreich’s
Spinocerebellar Ataxia (see extensive reviews on Friedreich Ataxia by Timchenko
and Caskey (1999), Pandolfo (2008)). This autosomal recessive congenital disease
is known to be caused by a GAA triplet “run” in the first intron of the FXN
(originally, known as X25) gene on 9q13-q21 that codes for a protein frataxin.
This protein is essential for mitochondria, as in its absence iron builds up and causes
free radical damage in nerve cells (such as in the cerebellum) and in muscle cells –
that is often the cause of heart failure in those affected by Friedreich. It is
a particularly interesting case, since the GAA “run” is intronic, thus it does not
result in the production of abnormal frataxin proteins. Instead, the mutation in the
1406 A.J. Pellionisz et al.
regulatory sequence causes gene silencing (Castaldo et al. 2008). Thus, an insuffi-
cient amount of Frataxin – or in more serious cases, a long tract of GAA repeats,
structurally weakens the DNA strand and the chromosome through breakage, as
evidenced through in vivo yeast studies. While a characteristically genomic disease,
Friedreich Ataxia is on the verge of therapy (Marmolino and Acquaviva 2009).
For the reasons above as reported earlier (Pellionisz 2009a), a structural analysis
of fractal defects was performed using FractoSoft Miner of HolGenTech, Inc. As
shown in Fig. 61.9, the fractal defect disrupting regulatory function was found
below the GAA triplet repeat in the middle of an (intronic) Alu repeat (see PLE-s
displayed in various colors).
Examination of long (or full) DNA sequences for fractal defects is made
important by the logic that since the genome is fractal, the actual sequence must
obey the fractal laws for proper function. For about a dozen hereditary conditions
such fractal defects have been identified. This is promising also for a very important
practical-logistical reason. Our rapidly increasing tally of full DNA sequences
shows “structural variants,” how the individual genomes are different from one-
another. The different bases can be counted by millions. Therefore, a mere catalog-
ing of such variants is unlikely to be a solid strategy of hunting down diseases.
Some variants most likely only cause “human diversity.” Perhaps only a much
smaller set of variants could be the root causes for diseases. Mathematically
speaking, in the most famous fractal, the Mandelbrot-set (1983), the mind-boggling
“complexity” arises from the rather simple equation Z ¼ Z^2 + C. In the equation
C is a constant, that may have the value of c or D (etc.) and the fractal set still
emerges, and just looks somewhat different. The differences between individual
genomes, therefore, fall into two separate classes. “Structural variants” can be neatly
parsed into what we call “parametric structural variants” (PSV-s, e.g., various values
of the constant in the equation). The c, or D will not violate the pristine fractal
equations. However, human genomes are likely to harbor “syntax structural variants”
(SSV-s). These alterations can render the fractal equation invalid; Z 6¼ Z^2 + C, thus
the genome’s own fractality may be compromised (as with Friedreich Ataxia) or even
Fig. 61.9 Friedreich Spinocerebellar Ataxia is known to be caused by a GAA triplet repeat at
a known locus. Fractal analysis reveals a FractoSet of Pyknon-like elements (short oligos shown in
different colors). It is conspicuous that the fractal defect is disrupted by the GAA triplet repeat,
after Pellionisz (2009b); see detailed explanation there and also in the text below
61 Recursive Genome Function of the Cerebellum 1407
grossly violated (as with cancers). These “syntax structural variants” (SSV-s) can be
mathematically expected to be direct causes of genome mis-regulation.
Using a computer code metaphor to illuminate the above argument, an algorithm
can be implemented with harmless “structural variants of lines of code.” In these
cases, the versions of the code would all run, but perhaps some versions of the code
would more rapidly or slowly converge than others. However, if some lines of
code would contain syntax-errors, the code not only would never run, but could not
even be compiled. Beyond the above proof of concept with Friedreich, the
perspective of genomic cancer diagnosis looms, by means of Fractal Defect
Mining for SSV-s. This opportunity is further detailed in the “Cancer” see section
“Future Directions.”
Application of Fractal Genomics for CancerCancer is widely regarded as “the disease of the genome.” Scientific results abound
stating that the progression of genome mis-regulation causes massive amounts of
structural variants of the DNA (see recent reviews; Meyerson et al. (2011) and
Ozery-Flato et al. (2011)).
In cerebellar tumors, it was found that sonic hedgehog signaling regulates the
growth and patterning of the cerebellum (Dahmane and Ruiz i Altaba (1999)). Also,
retinoid-related orphan receptors (RORs) were found to play critical roles in cancer,
development, immunity, circadian rhythm, and cellular metabolism (Jetten 2009).
A link between RORg and cancer is emerging from studies showing increased
expression of Th17-associated genes, including (an at least three-component vec-
tor) RORg, IL-17, and IL-23.
A particularly strong study suggests a possible role for RORa in cancer devel-
opment (Jetten 2009). “The ROR a gene spans a 730 kb genomic region that islocated in the middle of the common fragile site FRA15A within chromosomal band15q22.2.... Common fragile sites are highly unstable genomic regions found in allindividuals and are hotspots for deletions and other genetic alterations that maylead to altered expression and function of genes encoded within these regions.Common fragile sites have been implicated in several human diseases and areassociated with a number of different cancer types . . . . Genomic instability withinFRA15A might lead to changes in the expression of RORa and play a role in thedevelopment of certain cancers. This hypothesis is consistent with observationsshowing that RORa mRNA expression is often down-regulated in tumor cell linesand primary cancer samples . . . Moreover, studies examining gene expressionprofiles in various cancers identified ROR a as a gene commonly down-regulatedin several tumor types, particularly breast and lung cancer . . . Analysis of themethylation status of a series of genes identified ROR a as one of methylation-silenced genes in gastric cancer cell lines (Yamashita et al. 2006).”
The latter is in agreement with the concept that reduced expression of RORaexpression positively correlates with tumor formation. A major factor for such
dramatic alterations appears to be the mis-regulation due to hypo-methylation of
DNA (Hansen et al. (2011)).
1408 A.J. Pellionisz et al.
In terms of the fractal iterative recursion of multigenic vectors through matrices,
the tentative diagram below illustrates the concept exposed for general audience
(Pellionisz 2008b, at minute 30:00). In Eigenstates, perused master switches are
methylated (in the diagram of Fig. 61.10 shown by white “cookie”) and DNA-
enhancer and suppressor vectors force fractal recursive iteration into next stage of
hierarchy. In the diagram, an erroneous methylation (shown by yellow “cookie”)
would result in perusing a master switch in an uncontrolled manner – thus, the
fractal growth of the neuron would degenerate into a proliferation, instead of
stopping at the full grown state of the cell.
It should be emphasized, that the tentative scheme shown in Fig. 61.10 is
a seminal concept where the diagram greatly simplifies a hypothetical cancerous
growth due to hypo-methylation of the genome. First, the diagram shows only
a Purkinje neuron, though it is observed by both Dahmane and Ruiz i Altaba
(1999) and Jetten (2009) that cerebellar cancers depend on an interaction of
Purkinje- and granule cells. Also, it should be pointed out that the recursive lines
between DNA regions and protein structures represent the action of not a “single-
gene to single-RNA, to single-protein” loop, but recursion of multicomponent
vectors; for instance, as shown in Jetten (2009) an at least three-component vector
(of RORg, IL-17, and IL-23).
Fig. 61.10 Genomic and epigenomic fractal iteration derailed (From minutes 30:00 of YouTube
“Is IT Ready for the Dreaded DNA Data Deluge”?). Further explanation is in the video Pellionisz
(2008b) at 30:00 min and in the text below
61 Recursive Genome Function of the Cerebellum 1409
From a flood of evidence it is clear, that development of the cerebellum requires
the multidimensional co-regulation of vectors of genes (Oberdick et al. 1993;
Barski et al. 2002). It might take substantial time to assemble a comprehensive
map of genic and regulatory sequences that result normal or pathological (e.g.,
cancerous) cerebellar neural networks.
Future Directions
Theory of Recursive Algorithms
The Principle of Recursive Genome Function peer-reviewed paper (Pellionisz
2008a), also disseminated for general audience (Pellionisz 2008b) and presented
for debate at Cold Spring Harbor Labs (Pellionisz 2009b), laid out an agenda also
in practical terms (Pellionisz 2010) calling for substantial time and resources.
As usual with a new set of principles, future directions abound beyond the
originally outlined boundaries. It is understood that the tasks for theory-
development outlined below will require substantial time, perhaps generations,
and sizable resources.
Neural Net Algorithms Comprise Massively Parallel and CoordinatedGenome FunctionA central thesis of this chapter is that both neuroscience and genomic are charac-
terized by the “many-to-many” concept, that has not been emphasized sufficiently in
the past of genomics. Neural Net algorithms, both the existing (see e.g., Anderson
et al. 1990) and the to-be-developed algorithms, are most likely to be deployed in the
analytics of massively parallel genome function. Much of the gene expression in
“coordinated genome function” of “single genes,” for example, in Operon-theory
(Jacob and Monod 1961), was based on a mind frame reminiscent of the “single
reflex loop” of early neuroscience and thus, coordinated genome function could not
gain as much ground for the past half of a century as it was inherent in their
initiative.
Integration of Neural Net and Fractal AlgorithmsThe principle that coordinated genome function is based on recursion puts enor-
mous emphasis on the accelerated development of the theory of recursive algo-
rithms suitable for an algorithmic (software-enabling) understanding of
coordinated genome function. In this regard, both Neural Net algorithms as well
as Fractal Geometry algorithms have to be much further developed and inte-
grated. While Fractal Iterative Recursion is already featured, this chapter empha-
sizes that the recursion is not a “single loop” but is implemented in a massively
parallel manner. Moreover, the genome is certainly not monofractal, but
multifractal, and thus algorithmic development must be directed accordingly
(Barnsley 2006).
1410 A.J. Pellionisz et al.
Develop and Integrate Quantum Theory of Neuroscience and GenomicsNeural firing of spikes and the A, C, T, and G bases of the DNA will similarly
require a recognition that science is facing a quantum system both in neuroscience
and genomics. The “aperiodical” covalent bindings predicted by the seminal idea of
Schr€odinger (1944) preceded the discovery of DNA bases that establish such
bindings – perhaps a reason why emergence of quantum theory is sluggish com-
pared to that of physics. It is a question, however, if the discrete units are A, C, T,
and G bases, or, rather the quanta are codons (both amino-acid coding, as well as
“pervasively transcribed” noncoding triplets), short repetitive motifs (but certainly
not arbitrarily picked three to eight character “words”), or fractal “pyknon-like
elements” (PLE-s).
In a theoretical unification, the question will arise if in Recursive Genome
Function: Contravariant (Amino-Acid producing genic vectors) and Covariant
(Protein-bonding DNA-site vectors) converge and thus, obey the Fractal Weyl’s
Law on Fractal Quantum Eigenstate (Shepelyansky (2008), see the original
Weyl’s Law (1912)).
Public Domain Agenda in Industrialization of Genomics: Local andGlobal Fractal Dimension as a Standard Definition for OptimallyDistinguishing Cancerous and Control Genomes Based on TheirAbstract Measure of “Roughness”
Given the conclusion that “the genome is fractal” there is an immediate need, to
accomplish by a common and publicly available standard, worked out by the joint
effort of all concerned (genome informatics firms, cancer- and genome centers,
etc.). The goal is to arrive at a commonly accepted best performing definition of the
global and local “abstract roughness” (fractal dimension), in a manner optimized
for detection of mis-regulated (cancerous) genomes by bringing out the difference
in fractality of cancerous and control DNA.
Genomics and the “New War on Cancer” (Watson 2008, 2009) could greatly
benefit from a common focused effort of leading mathematically minded genomists
devoted to this vital practical problem of postmodern genomics.
Fractal dimension of physical objects, normally in two- or three-dimensional
spaces can follow the definition based on how fully the object fills the available
space. For instance, the Hilbert-curve, shown on the Science cover article by
Erez-Lieberman et al. (2009) elaborating on the seminal concept of Grosberg
et al. (1988, 1993) shows the fractal folding of DNA – squeezing a 2-m-long double
helix into the 6-mm diameter nucleus of a cell; where the Hilbert-curve is not only
“knot-free” in order to ensure uninterrupted transcription, but is also ultra-dense,
that is, “space-filling” with the physical fractal dimension of 3.
It needs to be pointed out, that “fractal dimension” can be defined not only
for actual physical objects, but the “roughness” of the double helix (say, if you
would run through the thread your fingers equipped to feel, like a brail pattern,
61 Recursive Genome Function of the Cerebellum 1411
the A, C, T, and G bases separately) can also be measured, given that both the
“abstract object” and the “abstract embedding space” are appropriately defined.
In the past, there were several attempts at defining “DNA fractal dimension.”
Berthelsen et al. (1992, see their Fig. 4) used both a two-dimensional embedding in
a space spun by AT horizontal and CG vertical axes, as well as a four-dimensional
embedding in a space spun by the AA:TT horizontal, CC:GG vertical, AG:GA-GT:
TG and AC:CA-GT:TG diagonal axes. The Grosberg-school of fractal DNA,
beyond their seminal concept of fractal folding of DNA (Grosberg et al. 1988,
1993) also revisited the issue of fractality of DNA texts (Borovik et al. 1994).
The numerous early DNA fractal dimension studies were triggered by
Mandelbrot (1983) but were conducted much before the now multiple supporting
facts were available that both the genome is fractal (see section “The Genome is
Fractal! Proof of Concept and the Basis of Generalization: Whole Genome Analysis
Reveals Repetitive Motifs Conforming to the Zipf-Mandelbrot Parabolic Fractal
Distribution Law of the Frequency/Ranking Diagram” of this chapter, the entire
double helix folds in a fractal manner (Erez-Lieberman 2009)), and brain cells, such
as the Purkinje neurons, are fractal (Pellionisz 1989) – plus our novel explosive set of
data that not only the surface of cancerous cells differs in spatial fractal dimension
from the control cells (Dokukin et al. 2011), but rather, there is a massive
rearrangement in the structure (obviously affecting the local and global “roughness”)
of cancerous genomes, see a recent finding of Copy Number Variation "fractal
defect" as a root-source of cancer, by clogging the transparency of the fractal 3D
Hilbert-curve (Fundenberg et al. 2011).
Given the amount of rapidly amassed data of cancerous and control full human
DNA, it is an urgent as well as an eminently feasible project to arrive at the
definition of both “the abstract DNA roughness” as well as the “abstract space in
which it is embedded” with the definitions optimized for distinguishing cancerous
genomes from their pristine (control) sequences.
In fractal theory, objects can be measured by different standards (“yardsticks”).
The famous question “How Long Is the Coast of Britain?” by Mandelbrot (1967)
can be answered in an infinite number of ways – as the length minimal or infinite –
depending on how science defines the “yardstick” with which the same object is to
be measured. Likewise, in defining the abstraction of global and local “roughness”
of the genome, appropriately embedded into an abstract multidimensional space, it is
reasonable to expect that cancerous deterioration can be tracked by “measurement of
local and global fractal dimension,” thus providing a diagnostic tool – before
unregulated/malformed proteins appear as the result of genomic rearrangements.
By what yardstick does Industrialization of Genomics (starting with present
R&D of Cancer) best measure the fractal difference characteristic to cancerous
DNA (fragments)? While “fractal dimension” mathematical literature is rich,
genomic/methylomic data is only presently available to identify the most suitable
mathematical definition for this novel, but life-or-death application.
While earlier attempts focused on A, C, T, and G bases to define an abstrac-
tion (embedded either into a two-dimensional, or four-dimensional abstract space,
spun over nucleotides), novel research points into the possibility of defining an
1412 A.J. Pellionisz et al.
abstract space of codons (Perez 2011a), wherein both “protein-coding codons”
and the “pervasively transcribed” the so-called “noncoding triplets” would
also be embedded. Further considerations include methylation and chromatin
modulation – rendering the segments of DNA “unreadable,” either temporarily or
permanently. It is a matter of definition of an unreadable (silent) DNA segment that
is totally smooth (with fractal dimension zero) – or to the contrary, like an unpaved
terrain, “infinitely rough,” thus impossible to be traveled. Another matter of defi-
nition is in what abstract space are the abstract objects embedded. In codon-space
embedding, or pyknon-space embedding, the measures are not only numerically
different, but they are likely to bring out the differences in fractality of cancerous
and control DNA in a more-or-less revealing manner.
Presently there is enough public DNA (with control) of cancerous sequences, with
already plenty of evidence for massive pathological alterations. It is a task for
a community of leading experts to work out by what definitions we could get the
best standard to spot the fractal genomic alteration associated with the progression of
the disease. “The Fractal Yardstick for Cancer” will emerge as a public domain
accomplishment, yielding an optimized and standard definition for genome analytics.
Public domain DNA data are to be downloaded from Cancer Centers, worldwide.
With the body of fractal literature reviewed, box-counting and other available
algorithms will be critically applied to provide the best practical definition to bring
out differences in terms of the fractal dimension of DNA (entire or fragmental). It is
of particular significance that formerly DNA fractal dimension was not focusing on
the methylation of bases, though by rendering certain sequences unreadable the
fractal dimension of the retrievable DNA information is most certainly altered.
A community effort also provides the opportunity of running and re-running bench-
mark tests as the work of the study-group develops by cloud computing on the same
body of data. The initiative plans for deploying not only “public clouds” (composed
of serial computers). Because of considerations of human data privacy (HIPPA), later
deployment of proprietary algorithms by fiercely competitive Big Pharma (about as
unlikely to rely entirely on open-source of algos and software as financial computing
retains proprietary), this initiative proposes simultaneous deployment of “private
clouds”. These are composed of Hybrid computers are built for speed and physical
efficacy (footprint and energy conservation). A genome informatics specialist with
cross-disciplinary experience might be welcomed to lead this initiative. The group of
top experts is expected to define the mathematical and computing strategy.
Proprietary Agenda in Industrialization of Genomics
The “Battelle Report” (2011) sized up this May the Economic Impact of the Human
Genome Project; how $3.8 billion investment drove $796 billion in economic
impact, created 310,000 jobs, and launched the genomic revolution. Not unlike
how the development of the science of nuclear physics was a necessary but
unsatisfactory condition to develop nuclear industry, genome informatics should
be mindful that the industrialization of genomics might at any time become
61 Recursive Genome Function of the Cerebellum 1413
unsustainable unless the scientific challenge of understanding coordinated genome
function in an algorithmic software-enabling manner is met by an accelerated
agenda. The scientific challenge is complicated by the very beneficial involvement
of the private sector (global informatics and product companies, like Samsung,
Procter & Gamble, Nestle, Unilever, and of course global Pharma companies, like
Genentech/Roche, as well as private hospital systems with Cancer Centers in the
lead). Given the fact that there are about 1,000 Cancer Centers in the USA alone,
and over 400 cancer drugs involving practically all Pharma companies, as well as
the computerization of both the hardware and software of hospital systems being
a lucrative business, Industrialization of Genomics is likely to follow previous
complex models. Most notably, those of defense, financial computer science and
industry – with intertwining public and fiercely competitive thus strictly proprietary
intellectual property, based on in-house science.
Hybrid Computation on Private CloudsAs assessed recently, Schadt et al. (2010), Industrialization of Genomics enables
individual laboratories to affordably generate terabytes or even petabytes of data.
Fortunately, as pointed out in a general presentation (Pellionisz 2008b) “Is IT ready
for the Dreaded DNA Data Deluge,” the real challenge is not the readiness of
information technology, since earlier data-intensive applications (defense-, nuclear-,
financial-, meteorological-, graphic-computing, etc.) have all been dealt with
the immense computing industry. Thus, the main challenges are in Information
Theory as Genome Informatics is applied toward an algorithmic understanding of
genome-epigenome (hologenome) regulation. Some of the scientific algorithms, just
as in financial computing, are fiercely proprietary (not only to provide accurate
predictions, but to deliver them faster than the competition). Industrialization of
Genomics emerges with entire segments (biodefense, private-domain wellness, and
health care) in a proprietary fashion. Additionally, since genomics deals with human
data that are legally mandated (in the USA, by HIPPA) to be handled in a confidential
manner, not only algorithm-security, but data-security is also indispensable. Genome
Computing Architecture, therefore, emerges with special needs and solutions
(Pellionisz 2009b).
Thus, though global IT firms (Microsoft, Amazon, Google, Facebook, etc.) have
mastered handling petabytes by computing architectures distributed over massively
parallel systems, only the transient research and development phase of the Indus-
trialization of Genomics, when many volunteers forgo privacy for the interest of
faster progress, will permit the standard “public cloud computing.” Though public
clouds are increasingly more secure, for example, by encryption, Baylor at Hous-
ton, Texas, already decided that for genome computing the appropriate solution is
a “private cloud,” moreover a closed system that is composed by the hybrid (serial/
parallel) computers. These platforms, available off-the-shelf from many companies
with a highly successful record of applying them in defense computing, financial
computing, etc., additionally provide, for example for hospitals, the small foot-
print, low-energy-consumption advantages, and most of all the speed that will be
required for hospital applications, when a biopsy tissue-sample will have to be sent
1414 A.J. Pellionisz et al.
to the local computing lab and results could be relayed back to the operating theater
while the patient is still on the table. Sequencing and Analytics all performed
locally, fast and affordably, without shipping hard-disks or uploading and
downloading data.
Consumer Genomics in Continuous Customer CareIndustrialization of Genomics started with Consumer Genomics that the FDA of the
USA does not regulate 23andMe, Inc. and Navigenics, Inc. While the market of
health care (genomic diagnosis, pharmaco-genomics of patients) is likely to be
restrained by regulatory industries, there is already a global trend, both in Europe
and Asia, to extend the benefits of genome interrogation, genome sequencing, and
genome analytics to vast masses of consumers (Pellionisz (2010) “Shop for Your
Life – HolGenTech at PMWC2010”). This trend will switch one-time analytics into
continuous customer care, known in business as the most lucrative “repeat customer
mode.” The announcement of Samsung, starting to provide genome analytics by
September 1, 2011, signaled a change of times. A Genome-Based Economy has
already commenced.
Acknowledgments Upon presentation, an advice was received from Dr. Hamilton O’ Smith
(Venter Institute), to run the Zipf-Mandelbrot Fractal Parabolic Distribution Curve-test, as
a control, with an identical number of A, C, T, and G-s, randomly generated. Not only the
control-result did not show the Curve, but generated zero repetition for the identical overall length
and motif-requirements (there were no points to compose any curve). The creative suggestion is
gratefully acknowledged. The authors thank Paul Shapshak Ph.D., Division of Infectious Disease
and International Medicine and Dept. of Psychiatry and Behavioral Medicine, University of South
Florida College of Medicine, Tampa, Florida, to help prepare the manuscript and to Prof. Sergey
Petoukhov, Academician, Moscow, for reference to Gazale and appreciative comments of the
chapter on dual valence, the RNA system serving as a Genomic Cerebellum.
One of us (AJP) also gratefully acknowledges Prof. E.G. Rajan for awarding this work by
the “Distinguished Scientist” honor for the presentation of the concepts at the ICSCI 2012
International Conference on Systemics, Cybernetics and Informatics, Hyderabad, India.
References
Albus JS (1971) Theory of cerebellar function. Math Biosci 10(1/2):25–61
Amari S (1991) Dualistic geometry of the manifold of higher-order neurons. Neural Netw
4(4):443–451
Anderson JA (1990) 351–355. In: Anderson JA, Pellionisz A, Rosenfeld E Neurocomputing II.
Directions of research. MIT Press. http://usa-siliconvalley.com/inst/pellionisz/90_anderson/
90_anderson.html
Anderson JA, Pellionisz A, Rosenfeld E (1990) Neurocomputing II. Directions of research. MIT
Press, Cambridge, MA
Arneth BM (2010) Sequence variability and sequence evolution: An explanation of molecular
polymorphisms and why many molecular structures can be preserved although they are not
predominant. DNA Cell Biol 29(10):571–576. doi:10.1089/dna.2009.0942
Baliga NS, Pan M, Goo YA, Yi EC, Goodlett DR, Dimitrov K, Shannon P, Aebersold R, Ng WV,
Hood L (2002) Coordinate regulation of energy transduction modules in Halobacterium sp.
analyzed by a global systems approach. Proc Natl Acad Sci USA 99(23):14913–14918
61 Recursive Genome Function of the Cerebellum 1415
Bailey TL, Gribskov M (1998) Methods and statistics for combining match scores (MEME,
MAST). J Comput Biol 5:211–221
Barnsley MF (2006) Superfractals. Cambridge University Press, Cambridge
Barski JJ, Lauth M, Meyer M (2002) Genetic targeting of cerebellar Purkinje cells: history, current
status and novel strategies. Cerebellum 1:111–118
Battelle Technology Partnership (2011) Economic impact of the human genome project. How a
$3.8 Bn investment drove $796 billion in economic impact, created 310,000 jobs and launched
the genomic revolution. http://www.battelle.org/publications/humangenomeproject.pdf
Bertalanffy L (1934) Untersuchungen €uber die Gesetzlichkeit des Wachstums. I. Allgemeine
Grundlagen der Theorie; mathematische und physiologische Gesetzlichkeiten des Wachstums
bei Wassertieren. Arch Entwicklungsmech 131:613–652
Berthelsen CL, Glazier JA, Skolnick MH (1992) Global fractal dimension of human DNA
sequences treated as pseudorandom walks. Phys Rev 45(12). http://biocomplexity.indiana.
edu/jglazier/docs/papers/20_DNA_Analysis.pdf
Bieberich E (1999) Structure in human consciousness: Fractal approach to the topology of the self
perceiving an outer world in an inner space. http://cogprints.org
Bieberich E (2011) Introduction to the fractality principle of consciousness and the sentyon
postulate. Cogn Comput. doi:10.1007/s12559-011-9104-5. http://www.springerlink.com/
content/964347x7251k0p36/
Bloedel JR, Tillery SI, Pellionisz AJ (1988) Experimental–theoretical analysis of the intrinsic
geometry of limb movements. Neurosci Abst 14:952
Bonhoeffer S, Herz AV, Boerlijst MC, Nee S, Nowak MA, May RM (1997) No signs of hidden
language in noncoding DNA. Phys Rev Lett 76(11):11
Borovik AS, Grosberg AY, Frank-Kamenetskii MD (1994) Fractality of DNA texts. J Biomol
Struct Dyn 12(3):655–669
Braitenberg V (1967) Is the cerebellar cortex a biological clock in the millisecond range? Prog
Brain Res 25:334–346
Cartieri FJ (2009) Darwinism and Lamarckism before and after Weisman: a historical, philosoph-
ical, and methodological analysis. University of Pittsburg, pp 1–54. http://etd.library.pitt.edu/
ETD/available/etd-05152009-110207/unrestricted/CartieriFJ_BPhil_ETD2009.pdf
Castaldo I, Pinelli M, Monticelli A, Acquaviva F, Giacchetti M, Filla A et al (2008) DNA
methylation in intron 1 of the frataxin gene is related to GAA repeat length and age of onset
in Friedreich ataxia patients. J Med Genet 45(12):808–812
Chatzidimitriou-Dreismann CA, Steffer RM, Larhammar D (1996) Lack of biological significance
in the “linguistic features” of noncoding DNA - a quantitative analysis. Nucleic Acids Res
24(9):1676–1681
Chiappelli F, Shapshak P, Commins D, Singer E, Minagar E, Oluwadara O et al (2008) Molecular
epigenetics, chromatin, and NeuroAIDS/HIV: immunopathological implications.
Bioinformation 3(1):47–52. http://www.usa-siliconvalley.com/chiappelli.pdf
Church GM (2005) The personal genome project. EMBO and Nature Pub. Group, Mol Syst Biol
1:30. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1681452/?tool¼pmcentrez
Churchland PS (1986) Neurophilosophy: toward a unified science of the mind-brain. MIT Press.
http://usa-siliconvalley.com/inst/pellionisz/85_churchland/85_churchland.html
Collins F (2007) New findings challenge established views on human genome. Nature. http://
www.genome.gov/25521554
Crick F (1970, seminal notion 1956) Central Dogma of Molecular Biology. Nature 227(8):561–563.
http://profiles.nlm.nih.gov/ps/access/SCBCCH.pdf, 1956: see http://www.junkdna.com
D’Angelo E, Mazzarello P, Prestori F, Mapelli J, Solinas S, Lombardo P, Cesana E, Gandolfi D,
Congi L (2010) The cerebellar network: from structure to function and dynamics. Brain Res
Rev 66:5–15 Published by Elsevier
Dahmane N, Ruiz i Altaba A (1999) Sonic hedgehog regulates the growth and patterning of
the cerebellum. Development 126:3089–3100. http://dev.biologists.org/content/126/14/3089.
full.pdf
1416 A.J. Pellionisz et al.
Daunicht W, Pellionisz A (1987) Spatial arrangement of the vestibular and the oculomotor system in
the rat. Brain Res 435:48–56. http://www.ncbi.nlm.nih.gov/pubmed/3501327?dopt¼Abstract
Descartes R (1629) Treatise of man. Prometheus Books (in English, 2003)
Dokukin ME, Guz NV, Gaikwad RM, Woodworth CD, Sokolov I (2011) Cell surface as fractal:
Normal and cancerous cervical cells demonstrate different fractal behavior of surface adhesion
maps at the nanoscale. Phys Rev Lett 107:028101
Dow RS, Moruzzi G (1958) The physiology and pathology of the cerebellum. Minnesota Univer-
sity Press, Minneapolis
Dupre J, Barnes SB (2008) Genomes and what to make of them. University of Chicago Press,
Chicago
Eccles J, Ito M, Szentagothai J (1967) The cerebellum as a neural machine. Springer, New York
Eckmiller R (1990) Advanced neural computers. North Holland. http://www.amazon.com/
Advanced-Neural-Computers-R-Eckmiller/dp/0444884009/ref¼sr_1_2?ie¼UTF8&qid¼1319
745602&sr¼8-2
Elnitski L, Piontkivska H, Welch JR (2011) In: Fedorov A, Fedorava L (eds) Advances in genomic
sequence analysis and pattern discovery, Chapter 3. Science, engineering and biology infor-
matics, vol 7. World Scientific, Singapore, pp 65–93
Finger S (1994) Origins of neuroscience: a history of explorations into brain function. Oxford
University Press, New York
Fiori S (2008) Lie-group-type neural system learning by manifold retractions. Neural Netw
21(10):1524–1529. http://202.114.89.42/resource/pdf/2174.pdf
Fire A, Xu S, Montgomery M, Kostas S, Driver S, Mello C (1998) Potent and specific genetic
interference by double-stranded RNA in Caenorhabditis elegans. Nature 391(6669):806–811.
doi:10.1038/35888 DOI:dx.doi.org. PMID 9486653
Flam F (1994) Hints of a language in junk DNA. Science 266:1320
Flourens MJP (1824) Recherces experimentalses fur les proprietes et les sonctions du systeme
nerveux dans les anomaux vertebres. Crevot, Paris
Foucher I, Montesinos ML, Volovitch M, Prochiantz A, Trembleau A (2003) Joint regulation of the
MAP1B promoter by HNF3b/Foxa2 and Engrailed is the result of a highly conserved mechanism
for direct interaction of homeoproteins and Fox transcription factors. Development 130:1867–
1876. doi:10.1242/dev.00414. http://dev.biologists.org/content/130/9/1867.full.pdf
Francis J (2008) Philosophy of mathematics. Global Vision Publishing House, New Delhi
Fraser CM, Gocayne JD, White O, Adams MD, Clayton RA, Fleischmann RD et al (1995) The
minimal gene complement of Mycoplasma genitalium. Science 270(5235):397–403 NCBI
ascension number NCBI L43927
Fundenberg G, Getz G, Meyerson M,Mirny L (2011) High-order chromatin architecture determines
the landscape of chromosomal alterations in cancer. Nature precedings, hdl:10101/
npre.2011.6356.1 http://precedings.nature.com/documents/6356/version/1/files/npre20116356-1.
Gardner M (1970) Mathematical games. The fantastic combinations of John Conway’s new
solitaire game “life”. Scientific American 223:120–123. http://ddi.cs.uni-potsdam.de/
HyFISCH/Produzieren/lis_projekt/proj_gamelife/ConwayScientificAmerican.htm
Gazale MJ (1999) Gnomonl: from pharaohs to fractals. Princeton Univ. Press http://books.google.
com/books?id=R0d76m-Be10C&printsec=copyright&source=gbs_pub_info_s&cad=3#v=one
page&q&f=false
Gibbs RA, Jeffrey Rogers J, Katze MG, Bumgarner R, Weinstock GM, Mardis ER et al
(2007) Evolutionary and biomedical insights from the rhesus macaque genome science
316(5822):222–234. http://www.sciencemag.org/content/316/5822/222.long
Gielen CCAM, van Zuylen EJ (1985) Coordination of arm muscles during flexion and supination:
application of the tensor analysis approach. Neuroscience 17:527–539. http://www.
sciencedirect.com/science/article/pii/030645228690028X
Glasner ME, Yen CC, Ekland EH, Bartel DP (2000) Recognition of nucleoside triphosphates
during RNA-Catalyzed primer extension. Biochemistry 39:15556–15562
61 Recursive Genome Function of the Cerebellum 1417
Grosberg AY, Nechaev SK, Shakhnovich EI (1988) The role of topological constraints in the
kinetics of collapse of macromolecules. J Phys France 49:2095–2100. http://hal.archives-
ouvertes.fr/docs/00/21/08/91/PDF/ajp-jphys_1988_49_12_2095_0.pdf
Grosberg A, Rabin Y, Havlin S, Neer A (1993) Crumpled globule model of the three-dimensional
structure of DNA. Europhys Lett 23:373–378. http://havlin.biu.ac.il/PS/scan189.pdf
Hansen KD, TimpW, Bravo HC, Sabunciyan, Langmead B, McDonald OG et al. (2011) Increased
methylation variation in epigenetic domains across cancer types. Nat Genet 43(8):768–775
Haussler D (1995) A generalized hidden Markov model for DNA parsing. Extended abstract of talk
for the workshop on gene-finding and gene structure prediction. University of Pennsylvania
Holmes G (1939) The cerebellum of man. Brain 62:1–30
Hood LR, Timp W, Bravo HC, Sabunciyan, Langmead B, McDonald OG et al. (2002) My life and
adventures integrating biology and technology. Commemorative lecture given when awarded
the 2002 Kyoto Prize in Advanced Technologies
Hopfield JJ (1982) Neural networks and physical systems with emergent collective computational
properties. Proc Natl Acad Sci (USA) 79:2554–2558
Izzo JA, Kim N, Elmetwaly S, Schlick T (2011) RAG: an update to the RNA-As-Graphs resource.
BMC Bioinformatics 12:219. http://www.biomedcentral.com/1471-2105/12/219
Jacob F, Monod JJ (1961) Genetic regulatory mechanisms in the synthesis of proteins. Mol Biol
3:318–356
Jansen J, Brodal A (1954) Aspects of cerebellar anatomy. The Wistar Institute of Anatomy and
Biology
Jensen KL, Styczynsk MPi, Rigoutsos I, Stephanopoulos GN (2006) A generic motif discovery
algorithm for sequential data (GEMODA). Bioinformatics 22(1):21–28
Jetten AM (2009) Retinoid-related orphan receptors (RORs): critical roles in development, immu-
nity, circadian rhythm, and cellular metabolism. Nuclear Receptor 7:1–32. doi:10.1621/
nrs.07003. http://www.nursa.org/retrieveFile.cfm?type¼NRS&file¼nrs07003.pdf
Kornberg A, Baker TA (1992) DNA replication. University Science Book, New York
Kuhn TS (1962) The structure of scientific revolutions, 1st edn. University of Chicago Press, Chicago
Kyle LJ, Styczynski MP, Rigoutsos I, Stephanopoulos GN (2006) A generic motif
discovery algorithm for sequential data. Bioinformatics 22(1):21–28. http://bioinformatics.
oxfordjournals.org/content/22/1/21.full.pdf
Laczko J, Pellionisz AJ, Peterson BW, Buchanan TS (1987) Multidimensional sensorimotor
“patterns” arising from a graphics-based tensorial model of the neck-motor system. In: Society
of Neuroscience Abstracts 13, vol 1, p 372
Laczko J, Pellionisz A, Jongen H, Gielen CCCM (1988) Computer modeling of human forelimb
muscle activation in multidimensional intrinsic coordinate frames. Soc Neurosci Absts 14–2:955
Lander ES, Linton LM, Birren B, Nusbaum C, Zody MC, Baldwin J et al (2001) Initial sequencing
and analysis of the human genome. Nature 409:860–921. http://www.nature.com/nature/jour
nal/v409/n6822/full/409860a0.html
Lestienne F, Liverneaux P, Pellionisz A (1988) Morpho-anatomy of sub-occipital muscles in
monkey: a tensor model of the musculo-skeletal head-neck system. Reunion Commune de la
Physiological Society et de l’Association des Physiologistes, France 2 July 1988. Proceedings
Physiological Society Journal of Physiology (Lond) 2P
Lieberman-Aiden E et al (2009) The comprehensive mapping of long-range interactions reveals
folding principles of the human genome. Science 326. doi:10.1126/science.1181369
Llinas R, Roy S (2009) The “prediction imperative” as the basis for self-awareness. Phil Trans R Soc B
(2009) 364:1301–1307. doi:10.1098/rstb.2008.0309. http://rstb.royalsocietypublishing.org/
content/364/1521/1301.full.pdf
Lorente De No R (1933) Vestibulo-ocular reflex arc. Arch Neurol Psychiat (Chicago) 30:245–291
Mandelbrot B (1967) How long is the coast of Britain? Statistical self-similarity and fractional
dimension. Science 5(156):3775, pp 636–638. doi:10.1126/science.156.3775.636. http://www.
sciencemag.org/content/156/3775/636.abstract
1418 A.J. Pellionisz et al.
Mandelbrot BB (1983) The fractal geometry of nature, 2nd edn. Freeman, New York
Manfred R, Roy S (2001) (Quantum) space-time as a statistical geometry of fuzzy lumps and the
connection with random metric spaces. Class Quantum Grav 18:3039
Mantegna RNSV, Buldyrev AL, Goldberger S, Havlin C, Peng K, Simons M, Stanley HE
(1994) Linguistic features of noncoding DNA sequences. Phys Rev Lett 73:3169–3172
Manto M (2008) The cerebellum, cerebellar disorders, and cerebellar research-two centuries of
discoveries. Cerebellum 7(4):505–516
Manto M, Marmolino D (2009) Cerebellar ataxias. Curr Opin Neurol 22(4): 419–429. http://www.
ncbi.nlm.nih.gov/pubmed/19421057
Marcer PJ (1992) Order and chaos in DNA – the Denis Guichard Prizewinner: Jean-Claude Perez.
In: Kibernetes 1992, 21(2):60–61. ISSN 0368-492X. http://www.emeraldinsight.com/journals.
htm?articleid¼1663901&show¼abstract
Mardis ER (2006) Anticipating the $1000 genome. Genome Biol 7(7):12. doi:10.1186/gb-2006-7-
7-112
Marmolino D, Acquaviva F (2009) Friedreich’s Ataxia: from the (GAA)n repeat mediated
silencing to new promising molecules for therapy. Cerebellum 8(3):245–259
Marr D (1969) A theory of cerebellar cortex. J Physiol 202:437–470
Marr D (1982) Vision: a computational Investigation into the human representation and processing
of visual information. Freeman, New York
Mattick JS (2001) Non-coding RNAs: the architects of eukaryotic complexity. EMBO reports
2(11):986–991. doi:10.1093/embo-reports/kve230. http://www.nature.com/embor/journal/v2/
n11/full/embor291.html
Mattick JS (2004) The hidden genetic code of complex organisms. Sci Am 291(4):60–67
Mattick JS (2005) The functional genomics of noncoding RNA. Science 309(5740):1527–1528.
http://www.sciencemag.org/content/309/5740/1527
McCulloch WS, Pitts WH (1943) A logical calculus of the ideas immanent in nervous activity.
Bull Math Biophys 5:115–133
Mendel JG (1866) Versuche €uber PflanzenhybridenVerhandlungen des naturforschenden
Vereines. In: Br€unn Bd. IV f€ur das Jahr, 1865 Abhandlungen:3–47. For the English translation,see: Druery CT and William Bateson (1901). “Experiments in plant hybridization”. J of the
Royal Horticultural Society 26:1–32. Retrieved 2009-10-09
Meyerson M, Gabriel S, Getz G (2011) Advances in understanding cancer genomes through
second-generation sequencing. Nature Rev Gen 11:685–696, doi:10.1038/nrg2841
Minsky M, Papert S (1969) Perceptrons:an introduction to computational geometry. MIT Press,
Cambridge, MA. ISBN:0 262 63022 2
Moruzzi G (1950) Problems in cerebellar physiology. C. C Thomas, Springfield
Neumann J (1958) The computer and the brain (Mrs. Hepsa Ely Silliman Memorial Lectures).
Yale University Press, New Haven/London, Second Edition 2000, with Introduction by
Patricia Churchland and Paul Churchland
Oberdick J, Schilling K, Smeyne RJ, Corbin JG, Bocchiaro C, Morgan JI (1993) Control of
segment-like patterns of gene expression in the mouse cerebellum. Neuron 10(6):1007–1018.
doi:10.1016/0896-6273(93)90050-2. http://www.ncbi.nlm.nih.gov/pubmed/8318226
Ohno S (1972) So much “Junk DNA” in our genome. Brookhaven Symp Biol (23):366–370. http://
www.junkdna.com/ohno.html
Oller JW (2010) The antithesis of entropy: Biosemiotic communication from genetics to human
language with special emphasis on the Iimmune systems. Entropy 12:631–705. doi:10.3390/
e12040631. http://www.mdpi.com/1099-4300/12/4/631/pdf
Ozery-Flato M, Linhart C, Trakhtenbrot L, Izraeli S, Shamir R (2011) Large-scale analysis of
chromosomal aberrations in cancer karyotypes reveals two distinct paths to aneuploidy.
Genome Biol 12:R61. doi:10.1186/gb-2011-12-6-r61
Pandolfo M (2008) Friedreich ataxia. Arch Neurol 65(10):1296–1303. http://archneur.ama-assn.
org/cgi/reprint/65/10/1296.pdf
61 Recursive Genome Function of the Cerebellum 1419
Pellionisz AJ (1984a) Coordination: a vector–matrix description of transformations of overcomplete
CNS coordinates and a tensorial solution using the Moore-Penrose generalized inverse. J Theor
Biol 110:353–375. http://www.junkdna.com/pellionisz_ coordination.pdf
Pellionisz A (1985a) Tensorial aspects of the multidimensional approach to the vestibulo-
oculomotor reflex and gaze. In: Berthoz A and Melvill-Jones G (eds) Reviews of oculomotor
research. I. Adaptive mechanisms in gaze control. Elsevier, Amsterdam, pp 281–296. http://
usa-siliconvalley.com/inst/pellionisz/berthoz/berthoz.html
Pellionisz AJ (1985b) Tensor network theory of the central nervous system and sensorimotor
modeling. In: Palm G, Aertsen A (eds) Brain theory. Springer, Berlin/Heidelberg/New York,
pp 121–145
Pellionisz A (1986) David Marr’s theory of the cerebellar cortex: a model in brain theory for the
“Galilean Combination of Simplification, Unification and Mathematization”. In: Palm G,
Aertsen A (eds) Brain theory. Springer, Berlin/Heidelberg/New York, pp 253–257. http://
www.junkdna.com/pellionisz_marr.pdf
Pellionisz A (1987) Tensor network theory of the central nervous system. In: Adelman G (ed)
Encyclopaedia of neuroscience. Birkh€auser, Boston, Basel, Stuttgart, pp 1196–1198. http://
usa-siliconvalley.com/inst/pellionisz/encyclopaedia/encyclopaedia.html
Pellionisz AJ (1989) Neural geometry: towards a fractal model of neurons. In: Cotterill RMJ (ed)
Models of brain function. Cambridge University Press, Cambridge, pp 453–464. http://www.
junkdna.com/pellionisz_fractal_purkinje_neuron_1989.pdf
Pellionisz A (2002) FractoGene: utility to use self-similar repetitions in the language-like genetic
information as fractal sets. US Patent Application, 1 Aug 2002. In: Simons MJ, Pellionisz AJ
(2006a), see also SF-gate, 22 Nov 2002. http://www.junkdna.com/plotkin.htm
Pellionisz A (2003) FractoGene: Design-tool for protein-based self-assembling nanostructures,
materials and application. Invited Keynote Lecture at Nano-Bio Technology Session. In:
Proceedings of the 204th meeting of the electrochemical society, p 1195
Pellionisz A (2006) PostGenetics: the journey of discovering “junk DNA”; Genetics beyond
genes. Invited Keynote Lecture of European inaugural satellite symposium of international
postgenetics Society. In: Proceedings of international congress of immunogenomics and
immunomics, Akademiai Kiado, Budapest
Pellionisz A (2008a) The principle of recursive genome function. The Cerebellum 7(3):348–359.
doi:10.1007/s12311-008-0035-y. http://ww.junkdna.com/pellionisz_principle_of_recursive_
genome_function.pdf
Pellionisz A (2008b) Is it ready for the dreaded DNA data deluge? http://www.youtube.com/
watch?v¼WJMFuc75V_w
Pellionisz A (2009a) From the principle of recursive genome function to interpretation of
hologenome regulation by personal genome computers. Personal Genomes, Cold Spring
Harbor Laboratory, 14–17 Sept 2009. http://www.junkdna.com/pellionisz_csh.html
Pellionisz A (2009b) Personal genome computing: breakthroughs, risks and opportunities, Chur-
chill Club Panel. http://www.youtube.com/watch?v¼znowfg7WJ1Q
Pellionisz A (2010) Shop for your life – HolGenTech at PMWC2010. http://www.youtube.com/
watch?v¼mSRMCDCVg6Y
Pellionisz A, Graf W (1987) Tensor network model of the “three-neuron vestibulo-ocular reflex-arc.”
In: Cat J Theoretical Neurobiology 5:127–151. http://www.junkdna.com/pellionisz_graf_1987.pdf
Pellionisz AJ, Llinas R (1980) Tensorial approach to the geometry of brain function. Cerebellar
coordination via a metric tensor. Neuroscience 5:1761–1770. http://usa-siliconvalley.com/inst/
pellionisz/80_metric/80_metric.html
Pellionisz AJ, Llinas R (1985) Tensor network theory of the metaorganization of functional
geometries in the CNS. Neuroscience 16:245–273. http://www.junkdna.com/pellionisz_me
taorganization_1985.pdf
Pellionisz A, Szentagothai J (1973) Dynamic single unit simulation of a realistic cerebellar
network model. Brain Res 49:83–99, Elsevier. http://usa-siliconvalley.com/inst/pellionisz/
ap_szj_72/ap_szj_72.html
1420 A.J. Pellionisz et al.
Pellionisz AJ, LeGoff B, Laczko J, Berthoz A (1991) Multidimensional geometry intrinsic to head
movements around distributed centers of rotation: a neurocomputer paradigm. In: Berthoz A,
Graf W, Vidal P (eds) The head-neck sensory-motor system. Oxford University Press, pp 117–
125. http://books.google.com/books?id¼6gLEx5slpOQC&pg¼PA158&dq¼pellionisz+legoff
+laczko#v¼onepage&q&f¼false
Pellionisz AJ, Jorgensen CC, Werbos PJ (1992) Cerebellar neurocontroller project for aerospace
applications. In: IJCNN international joint conference on neural networks. IEEE Catalog
Number: 92CH3114-6 ISBN: Softbound Edition 0-7803-0559-0. http://usa-siliconvalley.
com/inst/pellionisz/ijcnn92/ijcnn92.html
Perez JC (1988a) De nouvelles voies vers l’Intelligence Artificielle: pluri-disciplinarite, auto-
organisation et reseaux neuronaux. Masson, Paris. ISBN 2-225-81815-0
Perez JC (1988b) Fractal chaos: a new neural network holographic model. In: INNS conference,
Neural Networks, Boston
Perez JC (1990a) Digital holograms computers, concepts and applications. In: Neural networks:
biological computers or electronic brains, Les entretiens de Lyon. Springer, ISBN 2-287-00051-8
Perez JC (1990b) Integers neural network systems (INNS) using resonance properties of
a Fibonacci’s chaotic golden neuron. Neural Netw 1:859–865. IEEE 90CH2879-5. INSPEC
Accession Number: 3926657. http://ieeexplore.ieee.org/Xplore/login.jsp?url¼http%3A%2F%
2Fieeexplore.ieee.org%2Fiel2%2F148%2F3745%2F00137678.pdf%3Farnumber%3D137678
&authDecision¼�203
Perez JC (1991) Chaos, DNA, and neuro-computers: a golden link: the hidden language of genes,
global language and order in the human genome. Specul Sci Technol 14:336–346
Perez JC (1997) L’ADN decrypte. Resurgence, Liege. ISBN 2-87211-017-8
Perez JC (2008) Scale invariance embedded votes and self-emerging binary logics in the whole
human genome. Relating the paper: what is complexity? by Gell-Mann PM. Complexity. 1(1)
1995, Wiley. http://www.scribd.com/doc/68476137/JcperezLiveBitsFromDNA-from-John-
NASH-Logics-to-J-C-PEREZ-Fractals-Biomathematics
Perez JC (2009a) Codex biogenesis. Resurgence, Liege. ISBN 2-87434-044-8. https://sites.google.
com/site/codexbiogenesis/, http://www.ei.ur.edu.uy/JCP_codex.pdf
Perez JC (2009b) Golden ratio and numbers in DNA. http://golden-ratio-in-dna.blogspot.com and
http://golden-ratio-in-dna.blogspot.com/2008/01/1997-perezs-generic-predictive-equation.html
Perez JC (2010) Codon population in single-stranded whole human genome DNA are fractal and fine-
tuned by the Golden Ration 1.618. Interdiscip Sci: Comput Life Sci 2(3):228–340. doi:10.1007/
s12539-010-0022-0. http://urbanshakedowns.files.wordpress.com/2011/01/adn-perez.pdf
Perez JC (2011a) Decoding non-coding DNA codes: human genomemeta-chromosomes architecture
(support from Pr Luc Montagnier FMPRS World AIDS Foundation UNESCO and Jean-rene
Fourtou Vivendi Universal chairman), BIT Life Sciences’ third annual world vaccine congress-
2011. Beijing March 2011. http://www.scribd.com/doc/57828784/jcperezBeijing032011
Perez JC (2011b) Caminos Interdisciplinaios, Seminario CLAVE_INTER, Espacio Interdisci-
plinario, Universidad de la Republica Montevideo Uruguay, 27 de Octubre 2011. http://
www.ei.ur.edu.uy/enclave.html#confperez
Perez JC (2012) Paper in preparation: DNA, waveforms and numbers: Unifying all 1496 HIV1
whole genomes
Perez JC, Bertille JM (1990) A spatio temporal novelty detector using Fractal Chaos model. In:
IJCNN conference, Neural Networks, Washington
Peterson BW, Baker JA, Pellionisz AJ (1987) Multidimensional analysis of vestibulo-ocular and
vestibulo-colllic reflexes (VOR and VCR). In: Proceedings of the international symposium on
basic and applied aspects of vestibular function, Hong Kong
Peterson BW, Pellionisz AJ, Baker JA, Keshner EA (1989) Functional morphology and neural
control of neck muscles in mammals. Am Zoology 29:139–149
Petoukhov S, He M (2010) Symmetrical analysis techniques for genetic systems and bioinformat-
ics: advanced patterns and applications. Medical Information Science Reference. ISBN 978–
1605661247. http://www.flipkart.com/books/1605661244
61 Recursive Genome Function of the Cerebellum 1421
Rigoutsos I (2006) BM Watson Research Center. http://cbcsrv.watson.ibm.com/Tspd.html
Rigoutsos I, Floratos A (1998) Combinatorial pattern discovery in biological sequences: the
TEIRESIAS algorithm. Bioinformatics 14(1):55–67, pmid:9520502
Rigoutsos I, Huynh T, Miranda K, Tsirigos A, McHardy A, Platt D (2006) Short blocks from the
noncoding parts of the human genome have instances within nearly all known genes and relate
to biological processes. Proc Natl Acad Sci 103(17):6605–6610. http://www.pnas.org/content/
103/17/6605.full
Rolando L (1908) A reference handbook of the medical sciences, Vol. 1. Free Google e-Book
http://books.google.com/books?id=ltcEPWDHIkMC&pg=PA313&dq=Luigi+Rolando+(1908)
&hl=en&sa=X&ei=arQMT-6ZN4Ty2QWN5vzYBw&ved=0CEQQ6AEwAw#v=onepage&q=
larrey&f=false
Roy S, Llinas R (2007) Dynamic geometry, brain function modeling, and consciousness. Prog
Brain Res. 168:133–144. doi:10.1016/S0079-6123(07)68011-X. http://www.sciencedirect.
com/science/article/pii/S007961230768011X
Schadt EE, Linderman MD, Sorenson J, Lee L, Nolan GP (2010) Computational solutions to large-
scale data management and analysis. Nature Rev Genet (11):647–amp;657. http://www.nature.
com/nrg/journal/v11/n9/full/nrg2857.html
Schrodinger E (1944) What is life? Dublin Institute for Advanced Studies at Trinity College,
Dublin, in Feb 1943. http://whatislife.stanford.edu/LoCo_files/What-is-Life.pdf
Shannon CE (1948) Amathematical theory of communication. Bell Syst Tech J 27:379–423, 623–656
Shapshak P, Chiappelli, F, Commins D, Singer E, Levine AJ, Somboonwit C, Minagar A,
Pellionisz, A (2008) Molecular epigenetics, chromatin, and NeuroAIDS/HIV: translational
implications. Bioinformation 3(1):53–57. PMCID: PMC2586134 http://www.usa-
siliconvalley.com/shapshak.pdf
Shepelyansky DL (2008) Fractal Weyl law for quantum fractal eigenstates. Physical Rev E
77:015202. http://www.quantware.ups-tlse.fr/dima/myrefs/my165.pdf
Simons M, Pellionisz A (2006a) Genomics, morphogenesis and biophysics: triangulation of
purkinje cell development. The Cerebellum 5(1):27–35. http://www.junkdna.com/fractogene/
05_simons_pellionisz.pdf
Simons M, Pellionisz A (2006b) Implications of fractal organization of DNA on disease risk
genomic mapping and immune function analysis. Australasian and Southeast Asian Tissue
Typing Association. In: 30th scientific meeting 22–24 Nov 2006, Chiangmai. http://www.
junkdna.com/fractogene/06_simons_pellionisz.html
Stagnaro S (2011) Glycocalix quantum-biophysical-semeiotic evaluation plays a central role in
demonstration of water memory-information. http://www.sisbq.org/uploads/5/6/8/7/5687930/
wmi_glycocalyx.pdf
Stagnaro S, Caramel S (2011) A new way of therapy based on water memory-information: the
Quantum biophysical approach. http://www.sisbq.org/uploads/5/6/8/7/5687930/qbtherapy.pdf
Sylvester JJ (1853) On a theory of the syzygetic relations of two rational integral functions,
Comprising an application to the theory of Sturm’s functions, and that of the greatest algebraical
common measure. Philos Trans R Soc Lond 143:407–548. doi:10.1098/rstl.1853.0018, see also
http://en.wikipedia.org/wiki/Covariant_vector and http://en.wikipedia.org/wiki/Covariance_
and_contravariance_(computer_science)
Szentagothai J (1949) The elementary vestibulo-ocular reflex arc. J Neurophysiol 13(6):395–407.
http://jn.physiology.org/content/13/6/395.long
Timchenko LT, Caskey CT (1999) Triplet repeat disorders: discussion of molecular mechanisms.
Cell Mol Life Sci 55(11):1432–47. http://www.ncbi.nlm.nih.gov/pubmed/10518991
Venter C (2010) Multiple personal genomes await. Nature 464:676–677. doi:10.1038/464676a;
Published online 31 March 2010. http://www.nature.com/nature/journal/v464/n7289/full/
464676a.html
Venter JC, Adams MD, Myers EW, Li PW, Mural RJ, Sutton GG et al (2001) Science
291(5507):1304–1351. doi:10.1126/science.1058040. http://www.sciencemag.org/content/291/
5507/1304.full
1422 A.J. Pellionisz et al.
Wang L, Brown SJ (2006) Bind N: a web-based tool for efficient prediction of DNA and RNA
binding sites in amino acid sequences. Nucleic Acids Res 34:W243–W248. http://en.
wikipedia.org/wiki/RNA_interference
Watson JD (2008) Editorial: the failed war on cancer, People against cancer. http://www.
peopleagainstcancer.com/pdfs/news/20080916n2.pdf
Watson JD (2009) To fight cancer, know the enemy. New York Times August 6, A29 http://www.
nytimes.com/2009/08/06/opinion/06watson.html?pagewanted¼all
Weyl H (1912) Das asymptotisce lerteihingsgesetz der Eigenwerte linearer partieller Differential-
gleichungen. Math Ann 71:441–479. See also in http://media.wiley.com/product_data/excerpt/
04/35274083/3527408304.pdf
Wiener N (1948) Cybernetics or control and communication in the cnimal and the machine.
Hermann & Cie Editeurs/The Technology Press/Wiley, Paris, Cambridge, MA, New York.
http://www.amazon.com/Cybernetics-Second-Control-Communication-Machine/dp/026273009X/
ref¼sr_1_1?ie¼UTF8&qid¼1319564993&sr¼8-1
Yamashita S, Tsujino Y, Moriguchi K, Tatematsu M, Ushijima T (2006) Chemical genomic screen-
ing for methylation-silenced genes in gastric cancer cell lines using 5-aza-2’-deoxycytidine
treatment and oligonucleotide microarray. Cancer Sci 97:64–71. http://onlinelibrary.wiley.com/
doi/10.1111/j.1349-7006.2006.00136.x/pdf
Zhuangzi (around the 4th century BC) Stanford encyclopedia of philosophy http://plato.stanford.
edu/entries/taoism/
Zipf GK (1949) Human behaviour and the principle of least-effort. Addison-Wesley, Cambridge,MA
61 Recursive Genome Function of the Cerebellum 1423
Top Related